Hugging Face's logo

Datasets:
deeplumiere
/
EEE_datastore

License:
Dataset card Files
xet
 Community
12
deeplumiere commited on
Commit
ddeecdc
·
verified ·
1 Parent(s): e320917

Add files using upload-large-folder tool

Browse files
This view is limited to 50 files because it contains too many changes.   See raw diff
Files changed (50) hide show
  1. data/vals_ai/Qwen/Qwen2.5-72B-Instruct-Turbo/1228bed5-6603-4226-9ae7-67392111a123.json +491 -0
  2. data/vals_ai/Qwen/Qwen2.5-72B-Instruct-Turbo/bbe0ba41-7558-41aa-b811-4e6443348b81.json +420 -0
  3. data/vals_ai/Qwen/Qwen2.5-7B-Instruct-Turbo/e92c17e6-1d7d-48ea-bd3e-4fccdbd54b52.json +420 -0
  4. data/vals_ai/cursor/composer-2.5/0c990dc9-a21a-46e2-9759-7251f0e275ba.json +96 -0
  5. data/vals_ai/cursor/composer-2.5/9b814b26-9c1b-45f3-8fed-4c2a925349f2.json +284 -0
  6. data/vals_ai/cursor/composer-2.5/ec0da620-ae48-43dd-b83f-208074d5f946.json +346 -0
  7. data/vals_ai/devin/swe-1-6-fast/06177019-a9ba-46d3-af73-7c6a80172c97.json +89 -0
  8. data/vals_ai/google/gemini-1.0-pro-002/dcdcf7e4-a2e8-4e91-97b4-595d6ddf4cb6.json +408 -0
  9. data/vals_ai/google/gemini-1.5-flash-001/f3f8b998-89e7-41a7-a284-b83544188b6a.json +276 -0
  10. data/vals_ai/google/gemini-1.5-flash-002/03b06b69-5f43-40ea-9cd2-21a044e0ca86.json +981 -0
  11. data/vals_ai/google/gemini-1.5-flash-002/15f5e37d-a115-4d47-ad22-5d511fd7cff7.json +225 -0
  12. data/vals_ai/google/gemini-1.5-flash-002/2850f115-e5ce-42ce-a5bf-8bcaf55a091a.json +216 -0
  13. data/vals_ai/google/gemini-1.5-flash-002/4ee05046-cf52-4301-a477-282ede77be12.json +204 -0
  14. data/vals_ai/google/gemini-1.5-flash-002/6672691c-ff0e-409f-a8b4-ef6ac0316ce9.json +216 -0
  15. data/vals_ai/google/gemini-1.5-flash-002/7e409542-e6e4-4bd0-87bc-2fbfe8e1f846.json +408 -0
  16. data/vals_ai/google/gemini-1.5-flash-002/84088ee9-a4a6-4cf2-8d1d-4711ab9ed15a.json +792 -0
  17. data/vals_ai/google/gemini-1.5-flash-002/8f75c4eb-0ceb-438e-89e2-fbac1d1b0424.json +288 -0
  18. data/vals_ai/google/gemini-1.5-flash-002/b4d5c708-7af5-437d-ab02-451fef628983.json +99 -0
  19. data/vals_ai/google/gemini-1.5-flash-002/bb25fb0d-5550-490b-9ebe-4cee8e8e73ea.json +276 -0
  20. data/vals_ai/google/gemini-1.5-flash-002/bbf2e577-cbd3-4231-919f-6fdfc402b4a4.json +99 -0
  21. data/vals_ai/google/gemini-1.5-pro-002/275a02e0-dccc-422d-8069-d6a521becf92.json +225 -0
  22. data/vals_ai/google/gemini-1.5-pro-002/4f2e817f-6b6d-48b2-8ce5-4be91e29d426.json +792 -0
  23. data/vals_ai/google/gemini-1.5-pro-002/7635f61f-03d8-4879-9f76-5c870a56b2b2.json +288 -0
  24. data/vals_ai/google/gemini-1.5-pro-002/7e5903c6-0243-4a70-9eee-75d4c2b19c9e.json +99 -0
  25. data/vals_ai/google/gemini-1.5-pro-002/82982b53-113b-44fd-847f-ddbb3bf8dd94.json +204 -0
  26. data/vals_ai/google/gemini-1.5-pro-002/8a615867-e648-4856-a8e0-33c86823304f.json +276 -0
  27. data/vals_ai/google/gemini-1.5-pro-002/8d504161-486f-4732-93cf-389cebc550d8.json +981 -0
  28. data/vals_ai/google/gemini-1.5-pro-002/aa3c796d-e4ab-4c30-ba04-797de0a30464.json +99 -0
  29. data/vals_ai/google/gemini-1.5-pro-002/abab70b0-c577-4964-93e3-7795f9c3928a.json +408 -0
  30. data/vals_ai/google/gemini-1.5-pro-002/ccc83c77-cdf6-4dff-a2af-8fdfd7994da0.json +477 -0
  31. data/vals_ai/google/gemini-1.5-pro-002/dfe50d40-073b-464f-a6bd-c9ad72b68c4d.json +216 -0
  32. data/vals_ai/google/gemini-1.5-pro-002/ff92616e-001f-403a-88d2-07ce2cf8cc95.json +216 -0
  33. data/vals_ai/google/gemini-2.0-flash-001/14c3feef-f7fc-49a4-b370-e10c78646af7.json +408 -0
  34. data/vals_ai/google/gemini-2.0-flash-001/32c87288-dbd0-4b0c-b10f-615110169253.json +288 -0
  35. data/vals_ai/google/gemini-2.0-flash-001/39454daa-ad55-452f-a9fd-30f95d80ddf8.json +216 -0
  36. data/vals_ai/google/gemini-2.0-flash-001/3f54213b-2b24-4a9c-a3d6-1c2c16a52060.json +225 -0
  37. data/vals_ai/google/gemini-2.0-flash-001/6780858a-b0a5-4d52-b273-a01126dd746c.json +99 -0
  38. data/vals_ai/google/gemini-2.0-flash-001/6bd9c901-bd97-4f47-9dda-811857eb95ab.json +216 -0
  39. data/vals_ai/google/gemini-2.0-flash-001/85735689-b835-41fe-ba96-c2987c64e3da.json +276 -0
  40. data/vals_ai/google/gemini-2.0-flash-001/9e750eb8-5989-42df-ab2e-89febc4e4e7c.json +792 -0
  41. data/vals_ai/google/gemini-2.0-flash-001/a98d1b8f-32ce-4bb4-a010-55f3e1af3612.json +981 -0
  42. data/vals_ai/google/gemini-2.0-flash-001/c14ecf63-06fc-4adf-897d-48a25349238e.json +477 -0
  43. data/vals_ai/google/gemini-2.0-flash-001/c87b3ec1-7a1d-4560-8645-85a49f688348.json +204 -0
  44. data/vals_ai/google/gemini-2.0-flash-001/fbc69c3d-519b-4f50-b641-b85d8c250f5a.json +99 -0
  45. data/vals_ai/google/gemini-2.0-flash-exp/0edd0e94-356f-467a-8eb2-d1731327eb0d.json +99 -0
  46. data/vals_ai/google/gemini-2.0-flash-exp/3fd0a48d-98f8-4252-a624-a90d89e91c93.json +216 -0
  47. data/vals_ai/google/gemini-2.0-flash-exp/95074e1a-b052-4a19-b965-a760dd8e3934.json +216 -0
  48. data/vals_ai/google/gemini-2.0-flash-thinking-exp-01-21/39409693-1438-4025-a6ff-41c8c695b98e.json +216 -0
  49. data/vals_ai/google/gemini-2.0-flash-thinking-exp-01-21/5af204ba-31b0-4966-b1d0-b0e8bc39ee46.json +99 -0
  50. data/vals_ai/google/gemini-2.0-pro-exp-02-05/17466b64-ee76-4942-9926-aabdbfa839f0.json +216 -0
data/vals_ai/Qwen/Qwen2.5-72B-Instruct-Turbo/1228bed5-6603-4226-9ae7-67392111a123.json ADDED
@@ -0,0 +1,491 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/medqa/Qwen_Qwen2.5-72B-Instruct-Turbo/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MedQA",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "medqa",
13
+ "benchmark_name": "MedQA",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "healthcare",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "Qwen2.5-72B-Instruct-Turbo",
27
+ "id": "Qwen/Qwen2.5-72B-Instruct-Turbo",
28
+ "developer": "Qwen",
29
+ "additional_details": {
30
+ "vals_model_id": "together/Qwen/Qwen2.5-72B-Instruct-Turbo",
31
+ "vals_provider": "Together AI"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "medqa:asian:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
37
+ "evaluation_name": "vals_ai.medqa.asian",
38
+ "source_data": {
39
+ "dataset_name": "MedQA - Asian",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/medqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "medqa",
46
+ "task_key": "asian",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Asian).",
53
+ "metric_id": "vals_ai.medqa.asian.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 74.6,
69
+ "details": {
70
+ "benchmark_slug": "medqa",
71
+ "benchmark_name": "MedQA",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "asian",
74
+ "task_name": "Asian",
75
+ "dataset_type": "public",
76
+ "industry": "healthcare",
77
+ "raw_score": "74.6",
78
+ "raw_stderr": "1.906",
79
+ "latency": "5.298",
80
+ "cost_per_test": "0.000844",
81
+ "temperature": "0.7",
82
+ "max_output_tokens": "4096",
83
+ "provider": "Together AI"
84
+ },
85
+ "uncertainty": {
86
+ "standard_error": {
87
+ "value": 1.906,
88
+ "method": "vals_reported"
89
+ }
90
+ }
91
+ },
92
+ "generation_config": {
93
+ "generation_args": {
94
+ "temperature": 0.7,
95
+ "max_tokens": 4096,
96
+ "max_attempts": 1
97
+ }
98
+ }
99
+ },
100
+ {
101
+ "evaluation_result_id": "medqa:black:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
102
+ "evaluation_name": "vals_ai.medqa.black",
103
+ "source_data": {
104
+ "dataset_name": "MedQA - Black",
105
+ "source_type": "url",
106
+ "url": [
107
+ "https://www.vals.ai/benchmarks/medqa"
108
+ ],
109
+ "additional_details": {
110
+ "benchmark_slug": "medqa",
111
+ "task_key": "black",
112
+ "dataset_type": "public",
113
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
114
+ }
115
+ },
116
+ "metric_config": {
117
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Black).",
118
+ "metric_id": "vals_ai.medqa.black.accuracy",
119
+ "metric_name": "Accuracy",
120
+ "metric_kind": "accuracy",
121
+ "metric_unit": "percent",
122
+ "lower_is_better": false,
123
+ "score_type": "continuous",
124
+ "min_score": 0.0,
125
+ "max_score": 100.0,
126
+ "additional_details": {
127
+ "score_scale": "percent_0_to_100",
128
+ "max_score_source": "fixed_percentage_bound",
129
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
130
+ }
131
+ },
132
+ "score_details": {
133
+ "score": 73.671,
134
+ "details": {
135
+ "benchmark_slug": "medqa",
136
+ "benchmark_name": "MedQA",
137
+ "benchmark_updated": "2026-04-16",
138
+ "task_key": "black",
139
+ "task_name": "Black",
140
+ "dataset_type": "public",
141
+ "industry": "healthcare",
142
+ "raw_score": "73.671",
143
+ "raw_stderr": "1.932",
144
+ "latency": "5.778",
145
+ "cost_per_test": "0.000853",
146
+ "temperature": "0.7",
147
+ "max_output_tokens": "4096",
148
+ "provider": "Together AI"
149
+ },
150
+ "uncertainty": {
151
+ "standard_error": {
152
+ "value": 1.932,
153
+ "method": "vals_reported"
154
+ }
155
+ }
156
+ },
157
+ "generation_config": {
158
+ "generation_args": {
159
+ "temperature": 0.7,
160
+ "max_tokens": 4096,
161
+ "max_attempts": 1
162
+ }
163
+ }
164
+ },
165
+ {
166
+ "evaluation_result_id": "medqa:hispanic:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
167
+ "evaluation_name": "vals_ai.medqa.hispanic",
168
+ "source_data": {
169
+ "dataset_name": "MedQA - Hispanic",
170
+ "source_type": "url",
171
+ "url": [
172
+ "https://www.vals.ai/benchmarks/medqa"
173
+ ],
174
+ "additional_details": {
175
+ "benchmark_slug": "medqa",
176
+ "task_key": "hispanic",
177
+ "dataset_type": "public",
178
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
179
+ }
180
+ },
181
+ "metric_config": {
182
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Hispanic).",
183
+ "metric_id": "vals_ai.medqa.hispanic.accuracy",
184
+ "metric_name": "Accuracy",
185
+ "metric_kind": "accuracy",
186
+ "metric_unit": "percent",
187
+ "lower_is_better": false,
188
+ "score_type": "continuous",
189
+ "min_score": 0.0,
190
+ "max_score": 100.0,
191
+ "additional_details": {
192
+ "score_scale": "percent_0_to_100",
193
+ "max_score_source": "fixed_percentage_bound",
194
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
195
+ }
196
+ },
197
+ "score_details": {
198
+ "score": 73.9,
199
+ "details": {
200
+ "benchmark_slug": "medqa",
201
+ "benchmark_name": "MedQA",
202
+ "benchmark_updated": "2026-04-16",
203
+ "task_key": "hispanic",
204
+ "task_name": "Hispanic",
205
+ "dataset_type": "public",
206
+ "industry": "healthcare",
207
+ "raw_score": "73.9",
208
+ "raw_stderr": "1.924",
209
+ "latency": "6.154",
210
+ "cost_per_test": "0.00085",
211
+ "temperature": "0.7",
212
+ "max_output_tokens": "4096",
213
+ "provider": "Together AI"
214
+ },
215
+ "uncertainty": {
216
+ "standard_error": {
217
+ "value": 1.924,
218
+ "method": "vals_reported"
219
+ }
220
+ }
221
+ },
222
+ "generation_config": {
223
+ "generation_args": {
224
+ "temperature": 0.7,
225
+ "max_tokens": 4096,
226
+ "max_attempts": 1
227
+ }
228
+ }
229
+ },
230
+ {
231
+ "evaluation_result_id": "medqa:indigenous:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
232
+ "evaluation_name": "vals_ai.medqa.indigenous",
233
+ "source_data": {
234
+ "dataset_name": "MedQA - Indigenous",
235
+ "source_type": "url",
236
+ "url": [
237
+ "https://www.vals.ai/benchmarks/medqa"
238
+ ],
239
+ "additional_details": {
240
+ "benchmark_slug": "medqa",
241
+ "task_key": "indigenous",
242
+ "dataset_type": "public",
243
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
244
+ }
245
+ },
246
+ "metric_config": {
247
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Indigenous).",
248
+ "metric_id": "vals_ai.medqa.indigenous.accuracy",
249
+ "metric_name": "Accuracy",
250
+ "metric_kind": "accuracy",
251
+ "metric_unit": "percent",
252
+ "lower_is_better": false,
253
+ "score_type": "continuous",
254
+ "min_score": 0.0,
255
+ "max_score": 100.0,
256
+ "additional_details": {
257
+ "score_scale": "percent_0_to_100",
258
+ "max_score_source": "fixed_percentage_bound",
259
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
260
+ }
261
+ },
262
+ "score_details": {
263
+ "score": 74.05,
264
+ "details": {
265
+ "benchmark_slug": "medqa",
266
+ "benchmark_name": "MedQA",
267
+ "benchmark_updated": "2026-04-16",
268
+ "task_key": "indigenous",
269
+ "task_name": "Indigenous",
270
+ "dataset_type": "public",
271
+ "industry": "healthcare",
272
+ "raw_score": "74.05",
273
+ "raw_stderr": "1.92",
274
+ "latency": "6.391",
275
+ "cost_per_test": "0.000852",
276
+ "temperature": "0.7",
277
+ "max_output_tokens": "4096",
278
+ "provider": "Together AI"
279
+ },
280
+ "uncertainty": {
281
+ "standard_error": {
282
+ "value": 1.92,
283
+ "method": "vals_reported"
284
+ }
285
+ }
286
+ },
287
+ "generation_config": {
288
+ "generation_args": {
289
+ "temperature": 0.7,
290
+ "max_tokens": 4096,
291
+ "max_attempts": 1
292
+ }
293
+ }
294
+ },
295
+ {
296
+ "evaluation_result_id": "medqa:overall:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
297
+ "evaluation_name": "vals_ai.medqa.overall",
298
+ "source_data": {
299
+ "dataset_name": "MedQA - Overall",
300
+ "source_type": "url",
301
+ "url": [
302
+ "https://www.vals.ai/benchmarks/medqa"
303
+ ],
304
+ "additional_details": {
305
+ "benchmark_slug": "medqa",
306
+ "task_key": "overall",
307
+ "dataset_type": "public",
308
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
309
+ }
310
+ },
311
+ "metric_config": {
312
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Overall).",
313
+ "metric_id": "vals_ai.medqa.overall.accuracy",
314
+ "metric_name": "Accuracy",
315
+ "metric_kind": "accuracy",
316
+ "metric_unit": "percent",
317
+ "lower_is_better": false,
318
+ "score_type": "continuous",
319
+ "min_score": 0.0,
320
+ "max_score": 100.0,
321
+ "additional_details": {
322
+ "score_scale": "percent_0_to_100",
323
+ "max_score_source": "fixed_percentage_bound",
324
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
325
+ }
326
+ },
327
+ "score_details": {
328
+ "score": 77.395,
329
+ "details": {
330
+ "benchmark_slug": "medqa",
331
+ "benchmark_name": "MedQA",
332
+ "benchmark_updated": "2026-04-16",
333
+ "task_key": "overall",
334
+ "task_name": "Overall",
335
+ "dataset_type": "public",
336
+ "industry": "healthcare",
337
+ "raw_score": "77.395",
338
+ "raw_stderr": "0.736",
339
+ "latency": "5.573",
340
+ "cost_per_test": "0.000836",
341
+ "temperature": "0.7",
342
+ "max_output_tokens": "4096",
343
+ "provider": "Together AI"
344
+ },
345
+ "uncertainty": {
346
+ "standard_error": {
347
+ "value": 0.736,
348
+ "method": "vals_reported"
349
+ }
350
+ }
351
+ },
352
+ "generation_config": {
353
+ "generation_args": {
354
+ "temperature": 0.7,
355
+ "max_tokens": 4096,
356
+ "max_attempts": 1
357
+ }
358
+ }
359
+ },
360
+ {
361
+ "evaluation_result_id": "medqa:unbiased:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
362
+ "evaluation_name": "vals_ai.medqa.unbiased",
363
+ "source_data": {
364
+ "dataset_name": "MedQA - Unbiased",
365
+ "source_type": "url",
366
+ "url": [
367
+ "https://www.vals.ai/benchmarks/medqa"
368
+ ],
369
+ "additional_details": {
370
+ "benchmark_slug": "medqa",
371
+ "task_key": "unbiased",
372
+ "dataset_type": "public",
373
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
374
+ }
375
+ },
376
+ "metric_config": {
377
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Unbiased).",
378
+ "metric_id": "vals_ai.medqa.unbiased.accuracy",
379
+ "metric_name": "Accuracy",
380
+ "metric_kind": "accuracy",
381
+ "metric_unit": "percent",
382
+ "lower_is_better": false,
383
+ "score_type": "continuous",
384
+ "min_score": 0.0,
385
+ "max_score": 100.0,
386
+ "additional_details": {
387
+ "score_scale": "percent_0_to_100",
388
+ "max_score_source": "fixed_percentage_bound",
389
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
390
+ }
391
+ },
392
+ "score_details": {
393
+ "score": 93.95,
394
+ "details": {
395
+ "benchmark_slug": "medqa",
396
+ "benchmark_name": "MedQA",
397
+ "benchmark_updated": "2026-04-16",
398
+ "task_key": "unbiased",
399
+ "task_name": "Unbiased",
400
+ "dataset_type": "public",
401
+ "industry": "healthcare",
402
+ "raw_score": "93.95",
403
+ "raw_stderr": "1.047",
404
+ "latency": "3.906",
405
+ "cost_per_test": "0.000768",
406
+ "temperature": "0.7",
407
+ "max_output_tokens": "4096",
408
+ "provider": "Together AI"
409
+ },
410
+ "uncertainty": {
411
+ "standard_error": {
412
+ "value": 1.047,
413
+ "method": "vals_reported"
414
+ }
415
+ }
416
+ },
417
+ "generation_config": {
418
+ "generation_args": {
419
+ "temperature": 0.7,
420
+ "max_tokens": 4096,
421
+ "max_attempts": 1
422
+ }
423
+ }
424
+ },
425
+ {
426
+ "evaluation_result_id": "medqa:white:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
427
+ "evaluation_name": "vals_ai.medqa.white",
428
+ "source_data": {
429
+ "dataset_name": "MedQA - White",
430
+ "source_type": "url",
431
+ "url": [
432
+ "https://www.vals.ai/benchmarks/medqa"
433
+ ],
434
+ "additional_details": {
435
+ "benchmark_slug": "medqa",
436
+ "task_key": "white",
437
+ "dataset_type": "public",
438
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
439
+ }
440
+ },
441
+ "metric_config": {
442
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (White).",
443
+ "metric_id": "vals_ai.medqa.white.accuracy",
444
+ "metric_name": "Accuracy",
445
+ "metric_kind": "accuracy",
446
+ "metric_unit": "percent",
447
+ "lower_is_better": false,
448
+ "score_type": "continuous",
449
+ "min_score": 0.0,
450
+ "max_score": 100.0,
451
+ "additional_details": {
452
+ "score_scale": "percent_0_to_100",
453
+ "max_score_source": "fixed_percentage_bound",
454
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
455
+ }
456
+ },
457
+ "score_details": {
458
+ "score": 74.2,
459
+ "details": {
460
+ "benchmark_slug": "medqa",
461
+ "benchmark_name": "MedQA",
462
+ "benchmark_updated": "2026-04-16",
463
+ "task_key": "white",
464
+ "task_name": "White",
465
+ "dataset_type": "public",
466
+ "industry": "healthcare",
467
+ "raw_score": "74.2",
468
+ "raw_stderr": "1.916",
469
+ "latency": "5.909",
470
+ "cost_per_test": "0.000851",
471
+ "temperature": "0.7",
472
+ "max_output_tokens": "4096",
473
+ "provider": "Together AI"
474
+ },
475
+ "uncertainty": {
476
+ "standard_error": {
477
+ "value": 1.916,
478
+ "method": "vals_reported"
479
+ }
480
+ }
481
+ },
482
+ "generation_config": {
483
+ "generation_args": {
484
+ "temperature": 0.7,
485
+ "max_tokens": 4096,
486
+ "max_attempts": 1
487
+ }
488
+ }
489
+ }
490
+ ]
491
+ }
data/vals_ai/Qwen/Qwen2.5-72B-Instruct-Turbo/bbe0ba41-7558-41aa-b811-4e6443348b81.json ADDED
@@ -0,0 +1,420 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/Qwen_Qwen2.5-72B-Instruct-Turbo/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "Qwen2.5-72B-Instruct-Turbo",
27
+ "id": "Qwen/Qwen2.5-72B-Instruct-Turbo",
28
+ "developer": "Qwen",
29
+ "additional_details": {
30
+ "vals_model_id": "together/Qwen/Qwen2.5-72B-Instruct-Turbo",
31
+ "vals_provider": "Together AI"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 83.012,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "83.012",
78
+ "raw_stderr": "1.026",
79
+ "latency": "0.513",
80
+ "temperature": "0.7",
81
+ "max_output_tokens": "4096",
82
+ "provider": "Together AI"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.026,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "temperature": 0.7,
94
+ "max_tokens": 4096,
95
+ "max_attempts": 1
96
+ }
97
+ }
98
+ },
99
+ {
100
+ "evaluation_result_id": "legal_bench:interpretation_tasks:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
101
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
102
+ "source_data": {
103
+ "dataset_name": "LegalBench - Interpretation Tasks",
104
+ "source_type": "url",
105
+ "url": [
106
+ "https://www.vals.ai/benchmarks/legal_bench"
107
+ ],
108
+ "additional_details": {
109
+ "benchmark_slug": "legal_bench",
110
+ "task_key": "interpretation_tasks",
111
+ "dataset_type": "public",
112
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
113
+ }
114
+ },
115
+ "metric_config": {
116
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
117
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
118
+ "metric_name": "Accuracy",
119
+ "metric_kind": "accuracy",
120
+ "metric_unit": "percent",
121
+ "lower_is_better": false,
122
+ "score_type": "continuous",
123
+ "min_score": 0.0,
124
+ "max_score": 100.0,
125
+ "additional_details": {
126
+ "score_scale": "percent_0_to_100",
127
+ "max_score_source": "fixed_percentage_bound",
128
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
129
+ }
130
+ },
131
+ "score_details": {
132
+ "score": 78.576,
133
+ "details": {
134
+ "benchmark_slug": "legal_bench",
135
+ "benchmark_name": "LegalBench",
136
+ "benchmark_updated": "2026-06-17",
137
+ "task_key": "interpretation_tasks",
138
+ "task_name": "Interpretation Tasks",
139
+ "dataset_type": "public",
140
+ "industry": "legal",
141
+ "raw_score": "78.576",
142
+ "raw_stderr": "0.2",
143
+ "latency": "0.592",
144
+ "temperature": "0.7",
145
+ "max_output_tokens": "4096",
146
+ "provider": "Together AI"
147
+ },
148
+ "uncertainty": {
149
+ "standard_error": {
150
+ "value": 0.2,
151
+ "method": "vals_reported"
152
+ }
153
+ }
154
+ },
155
+ "generation_config": {
156
+ "generation_args": {
157
+ "temperature": 0.7,
158
+ "max_tokens": 4096,
159
+ "max_attempts": 1
160
+ }
161
+ }
162
+ },
163
+ {
164
+ "evaluation_result_id": "legal_bench:issue_tasks:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
165
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
166
+ "source_data": {
167
+ "dataset_name": "LegalBench - Issue Tasks",
168
+ "source_type": "url",
169
+ "url": [
170
+ "https://www.vals.ai/benchmarks/legal_bench"
171
+ ],
172
+ "additional_details": {
173
+ "benchmark_slug": "legal_bench",
174
+ "task_key": "issue_tasks",
175
+ "dataset_type": "public",
176
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
177
+ }
178
+ },
179
+ "metric_config": {
180
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
181
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
182
+ "metric_name": "Accuracy",
183
+ "metric_kind": "accuracy",
184
+ "metric_unit": "percent",
185
+ "lower_is_better": false,
186
+ "score_type": "continuous",
187
+ "min_score": 0.0,
188
+ "max_score": 100.0,
189
+ "additional_details": {
190
+ "score_scale": "percent_0_to_100",
191
+ "max_score_source": "fixed_percentage_bound",
192
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
193
+ }
194
+ },
195
+ "score_details": {
196
+ "score": 82.193,
197
+ "details": {
198
+ "benchmark_slug": "legal_bench",
199
+ "benchmark_name": "LegalBench",
200
+ "benchmark_updated": "2026-06-17",
201
+ "task_key": "issue_tasks",
202
+ "task_name": "Issue Tasks",
203
+ "dataset_type": "public",
204
+ "industry": "legal",
205
+ "raw_score": "82.193",
206
+ "raw_stderr": "0.678",
207
+ "latency": "0.638",
208
+ "temperature": "0.7",
209
+ "max_output_tokens": "4096",
210
+ "provider": "Together AI"
211
+ },
212
+ "uncertainty": {
213
+ "standard_error": {
214
+ "value": 0.678,
215
+ "method": "vals_reported"
216
+ }
217
+ }
218
+ },
219
+ "generation_config": {
220
+ "generation_args": {
221
+ "temperature": 0.7,
222
+ "max_tokens": 4096,
223
+ "max_attempts": 1
224
+ }
225
+ }
226
+ },
227
+ {
228
+ "evaluation_result_id": "legal_bench:overall:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
229
+ "evaluation_name": "vals_ai.legal_bench.overall",
230
+ "source_data": {
231
+ "dataset_name": "LegalBench - Overall",
232
+ "source_type": "url",
233
+ "url": [
234
+ "https://www.vals.ai/benchmarks/legal_bench"
235
+ ],
236
+ "additional_details": {
237
+ "benchmark_slug": "legal_bench",
238
+ "task_key": "overall",
239
+ "dataset_type": "public",
240
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
241
+ }
242
+ },
243
+ "metric_config": {
244
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
245
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
246
+ "metric_name": "Accuracy",
247
+ "metric_kind": "accuracy",
248
+ "metric_unit": "percent",
249
+ "lower_is_better": false,
250
+ "score_type": "continuous",
251
+ "min_score": 0.0,
252
+ "max_score": 100.0,
253
+ "additional_details": {
254
+ "score_scale": "percent_0_to_100",
255
+ "max_score_source": "fixed_percentage_bound",
256
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
257
+ }
258
+ },
259
+ "score_details": {
260
+ "score": 79.403,
261
+ "details": {
262
+ "benchmark_slug": "legal_bench",
263
+ "benchmark_name": "LegalBench",
264
+ "benchmark_updated": "2026-06-17",
265
+ "task_key": "overall",
266
+ "task_name": "Overall",
267
+ "dataset_type": "public",
268
+ "industry": "legal",
269
+ "raw_score": "79.403",
270
+ "raw_stderr": "0.66",
271
+ "latency": "0.831",
272
+ "temperature": "0.7",
273
+ "max_output_tokens": "4096",
274
+ "provider": "Together AI"
275
+ },
276
+ "uncertainty": {
277
+ "standard_error": {
278
+ "value": 0.66,
279
+ "method": "vals_reported"
280
+ }
281
+ }
282
+ },
283
+ "generation_config": {
284
+ "generation_args": {
285
+ "temperature": 0.7,
286
+ "max_tokens": 4096,
287
+ "max_attempts": 1
288
+ }
289
+ }
290
+ },
291
+ {
292
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
293
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
294
+ "source_data": {
295
+ "dataset_name": "LegalBench - Rhetoric Tasks",
296
+ "source_type": "url",
297
+ "url": [
298
+ "https://www.vals.ai/benchmarks/legal_bench"
299
+ ],
300
+ "additional_details": {
301
+ "benchmark_slug": "legal_bench",
302
+ "task_key": "rhetoric_tasks",
303
+ "dataset_type": "public",
304
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
305
+ }
306
+ },
307
+ "metric_config": {
308
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
309
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
310
+ "metric_name": "Accuracy",
311
+ "metric_kind": "accuracy",
312
+ "metric_unit": "percent",
313
+ "lower_is_better": false,
314
+ "score_type": "continuous",
315
+ "min_score": 0.0,
316
+ "max_score": 100.0,
317
+ "additional_details": {
318
+ "score_scale": "percent_0_to_100",
319
+ "max_score_source": "fixed_percentage_bound",
320
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
321
+ }
322
+ },
323
+ "score_details": {
324
+ "score": 78.36,
325
+ "details": {
326
+ "benchmark_slug": "legal_bench",
327
+ "benchmark_name": "LegalBench",
328
+ "benchmark_updated": "2026-06-17",
329
+ "task_key": "rhetoric_tasks",
330
+ "task_name": "Rhetoric Tasks",
331
+ "dataset_type": "public",
332
+ "industry": "legal",
333
+ "raw_score": "78.36",
334
+ "raw_stderr": "1.173",
335
+ "latency": "0.545",
336
+ "temperature": "0.7",
337
+ "max_output_tokens": "4096",
338
+ "provider": "Together AI"
339
+ },
340
+ "uncertainty": {
341
+ "standard_error": {
342
+ "value": 1.173,
343
+ "method": "vals_reported"
344
+ }
345
+ }
346
+ },
347
+ "generation_config": {
348
+ "generation_args": {
349
+ "temperature": 0.7,
350
+ "max_tokens": 4096,
351
+ "max_attempts": 1
352
+ }
353
+ }
354
+ },
355
+ {
356
+ "evaluation_result_id": "legal_bench:rule_tasks:together/Qwen/Qwen2.5-72B-Instruct-Turbo:score",
357
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
358
+ "source_data": {
359
+ "dataset_name": "LegalBench - Rule Tasks",
360
+ "source_type": "url",
361
+ "url": [
362
+ "https://www.vals.ai/benchmarks/legal_bench"
363
+ ],
364
+ "additional_details": {
365
+ "benchmark_slug": "legal_bench",
366
+ "task_key": "rule_tasks",
367
+ "dataset_type": "public",
368
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
369
+ }
370
+ },
371
+ "metric_config": {
372
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
373
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
374
+ "metric_name": "Accuracy",
375
+ "metric_kind": "accuracy",
376
+ "metric_unit": "percent",
377
+ "lower_is_better": false,
378
+ "score_type": "continuous",
379
+ "min_score": 0.0,
380
+ "max_score": 100.0,
381
+ "additional_details": {
382
+ "score_scale": "percent_0_to_100",
383
+ "max_score_source": "fixed_percentage_bound",
384
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
385
+ }
386
+ },
387
+ "score_details": {
388
+ "score": 74.874,
389
+ "details": {
390
+ "benchmark_slug": "legal_bench",
391
+ "benchmark_name": "LegalBench",
392
+ "benchmark_updated": "2026-06-17",
393
+ "task_key": "rule_tasks",
394
+ "task_name": "Rule Tasks",
395
+ "dataset_type": "public",
396
+ "industry": "legal",
397
+ "raw_score": "74.874",
398
+ "raw_stderr": "2.819",
399
+ "latency": "1.868",
400
+ "temperature": "0.7",
401
+ "max_output_tokens": "4096",
402
+ "provider": "Together AI"
403
+ },
404
+ "uncertainty": {
405
+ "standard_error": {
406
+ "value": 2.819,
407
+ "method": "vals_reported"
408
+ }
409
+ }
410
+ },
411
+ "generation_config": {
412
+ "generation_args": {
413
+ "temperature": 0.7,
414
+ "max_tokens": 4096,
415
+ "max_attempts": 1
416
+ }
417
+ }
418
+ }
419
+ ]
420
+ }
data/vals_ai/Qwen/Qwen2.5-7B-Instruct-Turbo/e92c17e6-1d7d-48ea-bd3e-4fccdbd54b52.json ADDED
@@ -0,0 +1,420 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/Qwen_Qwen2.5-7B-Instruct-Turbo/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "Qwen2.5-7B-Instruct-Turbo",
27
+ "id": "Qwen/Qwen2.5-7B-Instruct-Turbo",
28
+ "developer": "Qwen",
29
+ "additional_details": {
30
+ "vals_model_id": "together/Qwen/Qwen2.5-7B-Instruct-Turbo",
31
+ "vals_provider": "Together AI"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 65.607,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "65.607",
78
+ "raw_stderr": "1.3",
79
+ "latency": "0.339",
80
+ "temperature": "0.7",
81
+ "max_output_tokens": "4096",
82
+ "provider": "Together AI"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.3,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "temperature": 0.7,
94
+ "max_tokens": 4096,
95
+ "max_attempts": 1
96
+ }
97
+ }
98
+ },
99
+ {
100
+ "evaluation_result_id": "legal_bench:interpretation_tasks:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
101
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
102
+ "source_data": {
103
+ "dataset_name": "LegalBench - Interpretation Tasks",
104
+ "source_type": "url",
105
+ "url": [
106
+ "https://www.vals.ai/benchmarks/legal_bench"
107
+ ],
108
+ "additional_details": {
109
+ "benchmark_slug": "legal_bench",
110
+ "task_key": "interpretation_tasks",
111
+ "dataset_type": "public",
112
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
113
+ }
114
+ },
115
+ "metric_config": {
116
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
117
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
118
+ "metric_name": "Accuracy",
119
+ "metric_kind": "accuracy",
120
+ "metric_unit": "percent",
121
+ "lower_is_better": false,
122
+ "score_type": "continuous",
123
+ "min_score": 0.0,
124
+ "max_score": 100.0,
125
+ "additional_details": {
126
+ "score_scale": "percent_0_to_100",
127
+ "max_score_source": "fixed_percentage_bound",
128
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
129
+ }
130
+ },
131
+ "score_details": {
132
+ "score": 73.348,
133
+ "details": {
134
+ "benchmark_slug": "legal_bench",
135
+ "benchmark_name": "LegalBench",
136
+ "benchmark_updated": "2026-06-17",
137
+ "task_key": "interpretation_tasks",
138
+ "task_name": "Interpretation Tasks",
139
+ "dataset_type": "public",
140
+ "industry": "legal",
141
+ "raw_score": "73.348",
142
+ "raw_stderr": "0.248",
143
+ "latency": "0.313",
144
+ "temperature": "0.7",
145
+ "max_output_tokens": "4096",
146
+ "provider": "Together AI"
147
+ },
148
+ "uncertainty": {
149
+ "standard_error": {
150
+ "value": 0.248,
151
+ "method": "vals_reported"
152
+ }
153
+ }
154
+ },
155
+ "generation_config": {
156
+ "generation_args": {
157
+ "temperature": 0.7,
158
+ "max_tokens": 4096,
159
+ "max_attempts": 1
160
+ }
161
+ }
162
+ },
163
+ {
164
+ "evaluation_result_id": "legal_bench:issue_tasks:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
165
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
166
+ "source_data": {
167
+ "dataset_name": "LegalBench - Issue Tasks",
168
+ "source_type": "url",
169
+ "url": [
170
+ "https://www.vals.ai/benchmarks/legal_bench"
171
+ ],
172
+ "additional_details": {
173
+ "benchmark_slug": "legal_bench",
174
+ "task_key": "issue_tasks",
175
+ "dataset_type": "public",
176
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
177
+ }
178
+ },
179
+ "metric_config": {
180
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
181
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
182
+ "metric_name": "Accuracy",
183
+ "metric_kind": "accuracy",
184
+ "metric_unit": "percent",
185
+ "lower_is_better": false,
186
+ "score_type": "continuous",
187
+ "min_score": 0.0,
188
+ "max_score": 100.0,
189
+ "additional_details": {
190
+ "score_scale": "percent_0_to_100",
191
+ "max_score_source": "fixed_percentage_bound",
192
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
193
+ }
194
+ },
195
+ "score_details": {
196
+ "score": 72.583,
197
+ "details": {
198
+ "benchmark_slug": "legal_bench",
199
+ "benchmark_name": "LegalBench",
200
+ "benchmark_updated": "2026-06-17",
201
+ "task_key": "issue_tasks",
202
+ "task_name": "Issue Tasks",
203
+ "dataset_type": "public",
204
+ "industry": "legal",
205
+ "raw_score": "72.583",
206
+ "raw_stderr": "0.822",
207
+ "latency": "0.324",
208
+ "temperature": "0.7",
209
+ "max_output_tokens": "4096",
210
+ "provider": "Together AI"
211
+ },
212
+ "uncertainty": {
213
+ "standard_error": {
214
+ "value": 0.822,
215
+ "method": "vals_reported"
216
+ }
217
+ }
218
+ },
219
+ "generation_config": {
220
+ "generation_args": {
221
+ "temperature": 0.7,
222
+ "max_tokens": 4096,
223
+ "max_attempts": 1
224
+ }
225
+ }
226
+ },
227
+ {
228
+ "evaluation_result_id": "legal_bench:overall:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
229
+ "evaluation_name": "vals_ai.legal_bench.overall",
230
+ "source_data": {
231
+ "dataset_name": "LegalBench - Overall",
232
+ "source_type": "url",
233
+ "url": [
234
+ "https://www.vals.ai/benchmarks/legal_bench"
235
+ ],
236
+ "additional_details": {
237
+ "benchmark_slug": "legal_bench",
238
+ "task_key": "overall",
239
+ "dataset_type": "public",
240
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
241
+ }
242
+ },
243
+ "metric_config": {
244
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
245
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
246
+ "metric_name": "Accuracy",
247
+ "metric_kind": "accuracy",
248
+ "metric_unit": "percent",
249
+ "lower_is_better": false,
250
+ "score_type": "continuous",
251
+ "min_score": 0.0,
252
+ "max_score": 100.0,
253
+ "additional_details": {
254
+ "score_scale": "percent_0_to_100",
255
+ "max_score_source": "fixed_percentage_bound",
256
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
257
+ }
258
+ },
259
+ "score_details": {
260
+ "score": 69.559,
261
+ "details": {
262
+ "benchmark_slug": "legal_bench",
263
+ "benchmark_name": "LegalBench",
264
+ "benchmark_updated": "2026-06-17",
265
+ "task_key": "overall",
266
+ "task_name": "Overall",
267
+ "dataset_type": "public",
268
+ "industry": "legal",
269
+ "raw_score": "69.559",
270
+ "raw_stderr": "0.783",
271
+ "latency": "0.395",
272
+ "temperature": "0.7",
273
+ "max_output_tokens": "4096",
274
+ "provider": "Together AI"
275
+ },
276
+ "uncertainty": {
277
+ "standard_error": {
278
+ "value": 0.783,
279
+ "method": "vals_reported"
280
+ }
281
+ }
282
+ },
283
+ "generation_config": {
284
+ "generation_args": {
285
+ "temperature": 0.7,
286
+ "max_tokens": 4096,
287
+ "max_attempts": 1
288
+ }
289
+ }
290
+ },
291
+ {
292
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
293
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
294
+ "source_data": {
295
+ "dataset_name": "LegalBench - Rhetoric Tasks",
296
+ "source_type": "url",
297
+ "url": [
298
+ "https://www.vals.ai/benchmarks/legal_bench"
299
+ ],
300
+ "additional_details": {
301
+ "benchmark_slug": "legal_bench",
302
+ "task_key": "rhetoric_tasks",
303
+ "dataset_type": "public",
304
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
305
+ }
306
+ },
307
+ "metric_config": {
308
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
309
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
310
+ "metric_name": "Accuracy",
311
+ "metric_kind": "accuracy",
312
+ "metric_unit": "percent",
313
+ "lower_is_better": false,
314
+ "score_type": "continuous",
315
+ "min_score": 0.0,
316
+ "max_score": 100.0,
317
+ "additional_details": {
318
+ "score_scale": "percent_0_to_100",
319
+ "max_score_source": "fixed_percentage_bound",
320
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
321
+ }
322
+ },
323
+ "score_details": {
324
+ "score": 72.957,
325
+ "details": {
326
+ "benchmark_slug": "legal_bench",
327
+ "benchmark_name": "LegalBench",
328
+ "benchmark_updated": "2026-06-17",
329
+ "task_key": "rhetoric_tasks",
330
+ "task_name": "Rhetoric Tasks",
331
+ "dataset_type": "public",
332
+ "industry": "legal",
333
+ "raw_score": "72.957",
334
+ "raw_stderr": "1.303",
335
+ "latency": "0.293",
336
+ "temperature": "0.7",
337
+ "max_output_tokens": "4096",
338
+ "provider": "Together AI"
339
+ },
340
+ "uncertainty": {
341
+ "standard_error": {
342
+ "value": 1.303,
343
+ "method": "vals_reported"
344
+ }
345
+ }
346
+ },
347
+ "generation_config": {
348
+ "generation_args": {
349
+ "temperature": 0.7,
350
+ "max_tokens": 4096,
351
+ "max_attempts": 1
352
+ }
353
+ }
354
+ },
355
+ {
356
+ "evaluation_result_id": "legal_bench:rule_tasks:together/Qwen/Qwen2.5-7B-Instruct-Turbo:score",
357
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
358
+ "source_data": {
359
+ "dataset_name": "LegalBench - Rule Tasks",
360
+ "source_type": "url",
361
+ "url": [
362
+ "https://www.vals.ai/benchmarks/legal_bench"
363
+ ],
364
+ "additional_details": {
365
+ "benchmark_slug": "legal_bench",
366
+ "task_key": "rule_tasks",
367
+ "dataset_type": "public",
368
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
369
+ }
370
+ },
371
+ "metric_config": {
372
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
373
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
374
+ "metric_name": "Accuracy",
375
+ "metric_kind": "accuracy",
376
+ "metric_unit": "percent",
377
+ "lower_is_better": false,
378
+ "score_type": "continuous",
379
+ "min_score": 0.0,
380
+ "max_score": 100.0,
381
+ "additional_details": {
382
+ "score_scale": "percent_0_to_100",
383
+ "max_score_source": "fixed_percentage_bound",
384
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
385
+ }
386
+ },
387
+ "score_details": {
388
+ "score": 63.301,
389
+ "details": {
390
+ "benchmark_slug": "legal_bench",
391
+ "benchmark_name": "LegalBench",
392
+ "benchmark_updated": "2026-06-17",
393
+ "task_key": "rule_tasks",
394
+ "task_name": "Rule Tasks",
395
+ "dataset_type": "public",
396
+ "industry": "legal",
397
+ "raw_score": "63.301",
398
+ "raw_stderr": "3.346",
399
+ "latency": "0.708",
400
+ "temperature": "0.7",
401
+ "max_output_tokens": "4096",
402
+ "provider": "Together AI"
403
+ },
404
+ "uncertainty": {
405
+ "standard_error": {
406
+ "value": 3.346,
407
+ "method": "vals_reported"
408
+ }
409
+ }
410
+ },
411
+ "generation_config": {
412
+ "generation_args": {
413
+ "temperature": 0.7,
414
+ "max_tokens": 4096,
415
+ "max_attempts": 1
416
+ }
417
+ }
418
+ }
419
+ ]
420
+ }
data/vals_ai/cursor/composer-2.5/0c990dc9-a21a-46e2-9759-7251f0e275ba.json ADDED
@@ -0,0 +1,96 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/vibe-code/cursor_composer-2.5/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - Vibe Code Bench v1.1",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "vibe-code",
13
+ "benchmark_name": "Vibe Code Bench v1.1",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "composer-2.5",
27
+ "id": "cursor/composer-2.5",
28
+ "developer": "cursor",
29
+ "additional_details": {
30
+ "vals_model_id": "cursor/composer-2.5",
31
+ "vals_provider": "Cursor"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "vibe-code:overall:cursor/composer-2.5:score",
37
+ "evaluation_name": "vals_ai.vibe-code.overall",
38
+ "source_data": {
39
+ "dataset_name": "Vibe Code Bench v1.1 - Overall",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "vibe-code",
43
+ "task_key": "overall",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for Vibe Code Bench v1.1 (Overall).",
50
+ "metric_id": "vals_ai.vibe-code.overall.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 49.614,
66
+ "details": {
67
+ "benchmark_slug": "vibe-code",
68
+ "benchmark_name": "Vibe Code Bench v1.1",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "overall",
71
+ "task_name": "Overall",
72
+ "dataset_type": "private",
73
+ "industry": "coding",
74
+ "raw_score": "49.614",
75
+ "raw_stderr": "4.972",
76
+ "latency": "592.9",
77
+ "cost_per_test": "0.502233",
78
+ "max_output_tokens": "200000",
79
+ "provider": "Cursor"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 4.972,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 200000,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ }
95
+ ]
96
+ }
data/vals_ai/cursor/composer-2.5/9b814b26-9c1b-45f3-8fed-4c2a925349f2.json ADDED
@@ -0,0 +1,284 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/terminal-bench-2-1/cursor_composer-2.5/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - Terminal-Bench 2.1",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "terminal-bench-2-1",
13
+ "benchmark_name": "Terminal-Bench 2.1",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "composer-2.5",
27
+ "id": "cursor/composer-2.5",
28
+ "developer": "cursor",
29
+ "additional_details": {
30
+ "vals_model_id": "cursor/composer-2.5",
31
+ "vals_provider": "Cursor"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "terminal-bench-2-1:easy:cursor/composer-2.5:score",
37
+ "evaluation_name": "vals_ai.terminal-bench-2-1.easy",
38
+ "source_data": {
39
+ "dataset_name": "Terminal-Bench 2.1 - Easy",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/terminal-bench-2-1"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "terminal-bench-2-1",
46
+ "task_key": "easy",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for Terminal-Bench 2.1 (Easy).",
53
+ "metric_id": "vals_ai.terminal-bench-2-1.easy.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 100.0,
69
+ "details": {
70
+ "benchmark_slug": "terminal-bench-2-1",
71
+ "benchmark_name": "Terminal-Bench 2.1",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "easy",
74
+ "task_name": "Easy",
75
+ "dataset_type": "public",
76
+ "industry": "coding",
77
+ "raw_score": "100",
78
+ "raw_stderr": "0",
79
+ "latency": "0",
80
+ "max_output_tokens": "200000",
81
+ "provider": "Cursor"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 0.0,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 200000,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "terminal-bench-2-1:hard:cursor/composer-2.5:score",
99
+ "evaluation_name": "vals_ai.terminal-bench-2-1.hard",
100
+ "source_data": {
101
+ "dataset_name": "Terminal-Bench 2.1 - Hard",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/terminal-bench-2-1"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "terminal-bench-2-1",
108
+ "task_key": "hard",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for Terminal-Bench 2.1 (Hard).",
115
+ "metric_id": "vals_ai.terminal-bench-2-1.hard.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 40.0,
131
+ "details": {
132
+ "benchmark_slug": "terminal-bench-2-1",
133
+ "benchmark_name": "Terminal-Bench 2.1",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "hard",
136
+ "task_name": "Hard",
137
+ "dataset_type": "public",
138
+ "industry": "coding",
139
+ "raw_score": "40",
140
+ "raw_stderr": "9.097",
141
+ "latency": "0",
142
+ "max_output_tokens": "200000",
143
+ "provider": "Cursor"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 9.097,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 200000,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "terminal-bench-2-1:medium:cursor/composer-2.5:score",
161
+ "evaluation_name": "vals_ai.terminal-bench-2-1.medium",
162
+ "source_data": {
163
+ "dataset_name": "Terminal-Bench 2.1 - Medium",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/terminal-bench-2-1"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "terminal-bench-2-1",
170
+ "task_key": "medium",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for Terminal-Bench 2.1 (Medium).",
177
+ "metric_id": "vals_ai.terminal-bench-2-1.medium.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 65.455,
193
+ "details": {
194
+ "benchmark_slug": "terminal-bench-2-1",
195
+ "benchmark_name": "Terminal-Bench 2.1",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "medium",
198
+ "task_name": "Medium",
199
+ "dataset_type": "public",
200
+ "industry": "coding",
201
+ "raw_score": "65.455",
202
+ "raw_stderr": "6.471",
203
+ "latency": "0",
204
+ "max_output_tokens": "200000",
205
+ "provider": "Cursor"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 6.471,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 200000,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "terminal-bench-2-1:overall:cursor/composer-2.5:score",
223
+ "evaluation_name": "vals_ai.terminal-bench-2-1.overall",
224
+ "source_data": {
225
+ "dataset_name": "Terminal-Bench 2.1 - Overall",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/terminal-bench-2-1"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "terminal-bench-2-1",
232
+ "task_key": "overall",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for Terminal-Bench 2.1 (Overall).",
239
+ "metric_id": "vals_ai.terminal-bench-2-1.overall.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/terminal-bench-2-1"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 58.427,
255
+ "details": {
256
+ "benchmark_slug": "terminal-bench-2-1",
257
+ "benchmark_name": "Terminal-Bench 2.1",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": "overall",
260
+ "task_name": "Overall",
261
+ "dataset_type": "public",
262
+ "industry": "coding",
263
+ "raw_score": "58.427",
264
+ "raw_stderr": "5.254",
265
+ "latency": "0",
266
+ "max_output_tokens": "200000",
267
+ "provider": "Cursor"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 5.254,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 200000,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ }
283
+ ]
284
+ }
data/vals_ai/cursor/composer-2.5/ec0da620-ae48-43dd-b83f-208074d5f946.json ADDED
@@ -0,0 +1,346 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/swebench/cursor_composer-2.5/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - SWE-bench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "swebench",
13
+ "benchmark_name": "SWE-bench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "composer-2.5",
27
+ "id": "cursor/composer-2.5",
28
+ "developer": "cursor",
29
+ "additional_details": {
30
+ "vals_model_id": "cursor/composer-2.5",
31
+ "vals_provider": "Cursor"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "swebench:1-4 hours:cursor/composer-2.5:score",
37
+ "evaluation_name": "vals_ai.swebench.1-4 hours",
38
+ "source_data": {
39
+ "dataset_name": "SWE-bench - 1-4 hours",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/swebench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "swebench",
46
+ "task_key": "1-4 hours",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for SWE-bench (1-4 hours).",
53
+ "metric_id": "vals_ai.swebench.1-4 hours.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 52.381,
69
+ "details": {
70
+ "benchmark_slug": "swebench",
71
+ "benchmark_name": "SWE-bench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "1-4 hours",
74
+ "task_name": "1-4 hours",
75
+ "dataset_type": "public",
76
+ "industry": "coding",
77
+ "raw_score": "52.381",
78
+ "raw_stderr": "7.8",
79
+ "latency": "0",
80
+ "max_output_tokens": "200000",
81
+ "provider": "Cursor"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 7.8,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 200000,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "swebench:15 min - 1 hour:cursor/composer-2.5:score",
99
+ "evaluation_name": "vals_ai.swebench.15 min - 1 hour",
100
+ "source_data": {
101
+ "dataset_name": "SWE-bench - 15 min - 1 hour",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/swebench"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "swebench",
108
+ "task_key": "15 min - 1 hour",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for SWE-bench (15 min - 1 hour).",
115
+ "metric_id": "vals_ai.swebench.15 min - 1 hour.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 78.927,
131
+ "details": {
132
+ "benchmark_slug": "swebench",
133
+ "benchmark_name": "SWE-bench",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "15 min - 1 hour",
136
+ "task_name": "15 min - 1 hour",
137
+ "dataset_type": "public",
138
+ "industry": "coding",
139
+ "raw_score": "78.927",
140
+ "raw_stderr": "2.529",
141
+ "latency": "0",
142
+ "max_output_tokens": "200000",
143
+ "provider": "Cursor"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 2.529,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 200000,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "swebench:<15 min fix:cursor/composer-2.5:score",
161
+ "evaluation_name": "vals_ai.swebench.<15 min fix",
162
+ "source_data": {
163
+ "dataset_name": "SWE-bench - <15 min fix",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/swebench"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "swebench",
170
+ "task_key": "<15 min fix",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for SWE-bench (<15 min fix).",
177
+ "metric_id": "vals_ai.swebench.<15 min fix.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 86.598,
193
+ "details": {
194
+ "benchmark_slug": "swebench",
195
+ "benchmark_name": "SWE-bench",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "<15 min fix",
198
+ "task_name": "<15 min fix",
199
+ "dataset_type": "public",
200
+ "industry": "coding",
201
+ "raw_score": "86.598",
202
+ "raw_stderr": "2.452",
203
+ "latency": "0",
204
+ "max_output_tokens": "200000",
205
+ "provider": "Cursor"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 2.452,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 200000,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "swebench:>4 hours:cursor/composer-2.5:score",
223
+ "evaluation_name": "vals_ai.swebench.>4 hours",
224
+ "source_data": {
225
+ "dataset_name": "SWE-bench - >4 hours",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/swebench"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "swebench",
232
+ "task_key": ">4 hours",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for SWE-bench (>4 hours).",
239
+ "metric_id": "vals_ai.swebench.>4 hours.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 66.667,
255
+ "details": {
256
+ "benchmark_slug": "swebench",
257
+ "benchmark_name": "SWE-bench",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": ">4 hours",
260
+ "task_name": ">4 hours",
261
+ "dataset_type": "public",
262
+ "industry": "coding",
263
+ "raw_score": "66.667",
264
+ "raw_stderr": "33.333",
265
+ "latency": "0",
266
+ "max_output_tokens": "200000",
267
+ "provider": "Cursor"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 33.333,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 200000,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ },
283
+ {
284
+ "evaluation_result_id": "swebench:overall:cursor/composer-2.5:score",
285
+ "evaluation_name": "vals_ai.swebench.overall",
286
+ "source_data": {
287
+ "dataset_name": "SWE-bench - Overall",
288
+ "source_type": "url",
289
+ "url": [
290
+ "https://www.vals.ai/benchmarks/swebench"
291
+ ],
292
+ "additional_details": {
293
+ "benchmark_slug": "swebench",
294
+ "task_key": "overall",
295
+ "dataset_type": "public",
296
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
297
+ }
298
+ },
299
+ "metric_config": {
300
+ "evaluation_description": "Accuracy reported by Vals.ai for SWE-bench (Overall).",
301
+ "metric_id": "vals_ai.swebench.overall.accuracy",
302
+ "metric_name": "Accuracy",
303
+ "metric_kind": "accuracy",
304
+ "metric_unit": "percent",
305
+ "lower_is_better": false,
306
+ "score_type": "continuous",
307
+ "min_score": 0.0,
308
+ "max_score": 100.0,
309
+ "additional_details": {
310
+ "score_scale": "percent_0_to_100",
311
+ "max_score_source": "fixed_percentage_bound",
312
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/swebench"
313
+ }
314
+ },
315
+ "score_details": {
316
+ "score": 79.6,
317
+ "details": {
318
+ "benchmark_slug": "swebench",
319
+ "benchmark_name": "SWE-bench",
320
+ "benchmark_updated": "2026-06-17",
321
+ "task_key": "overall",
322
+ "task_name": "Overall",
323
+ "dataset_type": "public",
324
+ "industry": "coding",
325
+ "raw_score": "79.6",
326
+ "raw_stderr": "1.804",
327
+ "latency": "0",
328
+ "max_output_tokens": "200000",
329
+ "provider": "Cursor"
330
+ },
331
+ "uncertainty": {
332
+ "standard_error": {
333
+ "value": 1.804,
334
+ "method": "vals_reported"
335
+ }
336
+ }
337
+ },
338
+ "generation_config": {
339
+ "generation_args": {
340
+ "max_tokens": 200000,
341
+ "max_attempts": 1
342
+ }
343
+ }
344
+ }
345
+ ]
346
+ }
data/vals_ai/devin/swe-1-6-fast/06177019-a9ba-46d3-af73-7c6a80172c97.json ADDED
@@ -0,0 +1,89 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/vibe-code/devin_swe-1-6-fast/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - Vibe Code Bench v1.1",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "vibe-code",
13
+ "benchmark_name": "Vibe Code Bench v1.1",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "swe-1-6-fast",
27
+ "id": "devin/swe-1-6-fast",
28
+ "developer": "devin",
29
+ "additional_details": {
30
+ "vals_model_id": "devin/swe-1-6-fast",
31
+ "vals_provider": "Devin"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "vibe-code:overall:devin/swe-1-6-fast:score",
37
+ "evaluation_name": "vals_ai.vibe-code.overall",
38
+ "source_data": {
39
+ "dataset_name": "Vibe Code Bench v1.1 - Overall",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "vibe-code",
43
+ "task_key": "overall",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for Vibe Code Bench v1.1 (Overall).",
50
+ "metric_id": "vals_ai.vibe-code.overall.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/vibe-code"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 7.22,
66
+ "details": {
67
+ "benchmark_slug": "vibe-code",
68
+ "benchmark_name": "Vibe Code Bench v1.1",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "overall",
71
+ "task_name": "Overall",
72
+ "dataset_type": "private",
73
+ "industry": "coding",
74
+ "raw_score": "7.22",
75
+ "raw_stderr": "2.635",
76
+ "latency": "189.132",
77
+ "max_output_tokens": "0",
78
+ "provider": "Devin"
79
+ },
80
+ "uncertainty": {
81
+ "standard_error": {
82
+ "value": 2.635,
83
+ "method": "vals_reported"
84
+ }
85
+ }
86
+ }
87
+ }
88
+ ]
89
+ }
data/vals_ai/google/gemini-1.0-pro-002/dcdcf7e4-a2e8-4e91-97b4-595d6ddf4cb6.json ADDED
@@ -0,0 +1,408 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/google_gemini-1.0-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.0-pro-002",
27
+ "id": "google/gemini-1.0-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.0-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:google/gemini-1.0-pro-002:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 74.505,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "74.505",
78
+ "raw_stderr": "1.106",
79
+ "latency": "0.386",
80
+ "max_output_tokens": "8192",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 1.106,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 8192,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "legal_bench:interpretation_tasks:google/gemini-1.0-pro-002:score",
99
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
100
+ "source_data": {
101
+ "dataset_name": "LegalBench - Interpretation Tasks",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/legal_bench"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "legal_bench",
108
+ "task_key": "interpretation_tasks",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
115
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 73.86,
131
+ "details": {
132
+ "benchmark_slug": "legal_bench",
133
+ "benchmark_name": "LegalBench",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "interpretation_tasks",
136
+ "task_name": "Interpretation Tasks",
137
+ "dataset_type": "public",
138
+ "industry": "legal",
139
+ "raw_score": "73.86",
140
+ "raw_stderr": "0.196",
141
+ "latency": "0.42",
142
+ "max_output_tokens": "8192",
143
+ "provider": "Google"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 0.196,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 8192,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "legal_bench:issue_tasks:google/gemini-1.0-pro-002:score",
161
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
162
+ "source_data": {
163
+ "dataset_name": "LegalBench - Issue Tasks",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/legal_bench"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "legal_bench",
170
+ "task_key": "issue_tasks",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
177
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 72.707,
193
+ "details": {
194
+ "benchmark_slug": "legal_bench",
195
+ "benchmark_name": "LegalBench",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "issue_tasks",
198
+ "task_name": "Issue Tasks",
199
+ "dataset_type": "public",
200
+ "industry": "legal",
201
+ "raw_score": "72.707",
202
+ "raw_stderr": "0.677",
203
+ "latency": "0.622",
204
+ "max_output_tokens": "8192",
205
+ "provider": "Google"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 0.677,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 8192,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "legal_bench:overall:google/gemini-1.0-pro-002:score",
223
+ "evaluation_name": "vals_ai.legal_bench.overall",
224
+ "source_data": {
225
+ "dataset_name": "LegalBench - Overall",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/legal_bench"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "legal_bench",
232
+ "task_key": "overall",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
239
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 70.247,
255
+ "details": {
256
+ "benchmark_slug": "legal_bench",
257
+ "benchmark_name": "LegalBench",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": "overall",
260
+ "task_name": "Overall",
261
+ "dataset_type": "public",
262
+ "industry": "legal",
263
+ "raw_score": "70.247",
264
+ "raw_stderr": "0.765",
265
+ "latency": "0.465",
266
+ "max_output_tokens": "8192",
267
+ "provider": "Google"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 0.765,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 8192,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ },
283
+ {
284
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:google/gemini-1.0-pro-002:score",
285
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
286
+ "source_data": {
287
+ "dataset_name": "LegalBench - Rhetoric Tasks",
288
+ "source_type": "url",
289
+ "url": [
290
+ "https://www.vals.ai/benchmarks/legal_bench"
291
+ ],
292
+ "additional_details": {
293
+ "benchmark_slug": "legal_bench",
294
+ "task_key": "rhetoric_tasks",
295
+ "dataset_type": "public",
296
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
297
+ }
298
+ },
299
+ "metric_config": {
300
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
301
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
302
+ "metric_name": "Accuracy",
303
+ "metric_kind": "accuracy",
304
+ "metric_unit": "percent",
305
+ "lower_is_better": false,
306
+ "score_type": "continuous",
307
+ "min_score": 0.0,
308
+ "max_score": 100.0,
309
+ "additional_details": {
310
+ "score_scale": "percent_0_to_100",
311
+ "max_score_source": "fixed_percentage_bound",
312
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
313
+ }
314
+ },
315
+ "score_details": {
316
+ "score": 77.78,
317
+ "details": {
318
+ "benchmark_slug": "legal_bench",
319
+ "benchmark_name": "LegalBench",
320
+ "benchmark_updated": "2026-06-17",
321
+ "task_key": "rhetoric_tasks",
322
+ "task_name": "Rhetoric Tasks",
323
+ "dataset_type": "public",
324
+ "industry": "legal",
325
+ "raw_score": "77.78",
326
+ "raw_stderr": "0.848",
327
+ "latency": "0.436",
328
+ "max_output_tokens": "8192",
329
+ "provider": "Google"
330
+ },
331
+ "uncertainty": {
332
+ "standard_error": {
333
+ "value": 0.848,
334
+ "method": "vals_reported"
335
+ }
336
+ }
337
+ },
338
+ "generation_config": {
339
+ "generation_args": {
340
+ "max_tokens": 8192,
341
+ "max_attempts": 1
342
+ }
343
+ }
344
+ },
345
+ {
346
+ "evaluation_result_id": "legal_bench:rule_tasks:google/gemini-1.0-pro-002:score",
347
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
348
+ "source_data": {
349
+ "dataset_name": "LegalBench - Rule Tasks",
350
+ "source_type": "url",
351
+ "url": [
352
+ "https://www.vals.ai/benchmarks/legal_bench"
353
+ ],
354
+ "additional_details": {
355
+ "benchmark_slug": "legal_bench",
356
+ "task_key": "rule_tasks",
357
+ "dataset_type": "public",
358
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
359
+ }
360
+ },
361
+ "metric_config": {
362
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
363
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
364
+ "metric_name": "Accuracy",
365
+ "metric_kind": "accuracy",
366
+ "metric_unit": "percent",
367
+ "lower_is_better": false,
368
+ "score_type": "continuous",
369
+ "min_score": 0.0,
370
+ "max_score": 100.0,
371
+ "additional_details": {
372
+ "score_scale": "percent_0_to_100",
373
+ "max_score_source": "fixed_percentage_bound",
374
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
375
+ }
376
+ },
377
+ "score_details": {
378
+ "score": 52.382,
379
+ "details": {
380
+ "benchmark_slug": "legal_bench",
381
+ "benchmark_name": "LegalBench",
382
+ "benchmark_updated": "2026-06-17",
383
+ "task_key": "rule_tasks",
384
+ "task_name": "Rule Tasks",
385
+ "dataset_type": "public",
386
+ "industry": "legal",
387
+ "raw_score": "52.382",
388
+ "raw_stderr": "3.493",
389
+ "latency": "0.461",
390
+ "max_output_tokens": "8192",
391
+ "provider": "Google"
392
+ },
393
+ "uncertainty": {
394
+ "standard_error": {
395
+ "value": 3.493,
396
+ "method": "vals_reported"
397
+ }
398
+ }
399
+ },
400
+ "generation_config": {
401
+ "generation_args": {
402
+ "max_tokens": 8192,
403
+ "max_attempts": 1
404
+ }
405
+ }
406
+ }
407
+ ]
408
+ }
data/vals_ai/google/gemini-1.5-flash-001/f3f8b998-89e7-41a7-a284-b83544188b6a.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/corp_fin_v2/google_gemini-1.5-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - CorpFin v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "corp_fin_v2",
13
+ "benchmark_name": "CorpFin v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-001",
27
+ "id": "google/gemini-1.5-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "corp_fin_v2:exact_pages:google/gemini-1.5-flash-001:score",
37
+ "evaluation_name": "vals_ai.corp_fin_v2.exact_pages",
38
+ "source_data": {
39
+ "dataset_name": "CorpFin v2 - Exact Pages",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "corp_fin_v2",
43
+ "task_key": "exact_pages",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Exact Pages).",
50
+ "metric_id": "vals_ai.corp_fin_v2.exact_pages.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 38.462,
66
+ "details": {
67
+ "benchmark_slug": "corp_fin_v2",
68
+ "benchmark_name": "CorpFin v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "exact_pages",
71
+ "task_name": "Exact Pages",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "38.462",
75
+ "raw_stderr": "1.661",
76
+ "latency": "3.353",
77
+ "cost_per_test": "0.00014",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.661,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "corp_fin_v2:max_fitting_context:google/gemini-1.5-flash-001:score",
97
+ "evaluation_name": "vals_ai.corp_fin_v2.max_fitting_context",
98
+ "source_data": {
99
+ "dataset_name": "CorpFin v2 - Max Fitting Context",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "corp_fin_v2",
103
+ "task_key": "max_fitting_context",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Max Fitting Context).",
110
+ "metric_id": "vals_ai.corp_fin_v2.max_fitting_context.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 25.991,
126
+ "details": {
127
+ "benchmark_slug": "corp_fin_v2",
128
+ "benchmark_name": "CorpFin v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "max_fitting_context",
131
+ "task_name": "Max Fitting Context",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "25.991",
135
+ "raw_stderr": "1.497",
136
+ "latency": "0.055",
137
+ "cost_per_test": "0.014092",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.497,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "corp_fin_v2:overall:google/gemini-1.5-flash-001:score",
157
+ "evaluation_name": "vals_ai.corp_fin_v2.overall",
158
+ "source_data": {
159
+ "dataset_name": "CorpFin v2 - Overall",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "corp_fin_v2",
163
+ "task_key": "overall",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Overall).",
170
+ "metric_id": "vals_ai.corp_fin_v2.overall.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 28.633,
186
+ "details": {
187
+ "benchmark_slug": "corp_fin_v2",
188
+ "benchmark_name": "CorpFin v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "overall",
191
+ "task_name": "Overall",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "28.633",
195
+ "raw_stderr": "0.88",
196
+ "latency": "1.156",
197
+ "cost_per_test": "0.006413",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 0.88,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ },
215
+ {
216
+ "evaluation_result_id": "corp_fin_v2:shared_max_context:google/gemini-1.5-flash-001:score",
217
+ "evaluation_name": "vals_ai.corp_fin_v2.shared_max_context",
218
+ "source_data": {
219
+ "dataset_name": "CorpFin v2 - Shared Max Context",
220
+ "source_type": "other",
221
+ "additional_details": {
222
+ "benchmark_slug": "corp_fin_v2",
223
+ "task_key": "shared_max_context",
224
+ "dataset_type": "private",
225
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
226
+ }
227
+ },
228
+ "metric_config": {
229
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Shared Max Context).",
230
+ "metric_id": "vals_ai.corp_fin_v2.shared_max_context.accuracy",
231
+ "metric_name": "Accuracy",
232
+ "metric_kind": "accuracy",
233
+ "metric_unit": "percent",
234
+ "lower_is_better": false,
235
+ "score_type": "continuous",
236
+ "min_score": 0.0,
237
+ "max_score": 100.0,
238
+ "additional_details": {
239
+ "score_scale": "percent_0_to_100",
240
+ "max_score_source": "fixed_percentage_bound",
241
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
242
+ }
243
+ },
244
+ "score_details": {
245
+ "score": 21.445,
246
+ "details": {
247
+ "benchmark_slug": "corp_fin_v2",
248
+ "benchmark_name": "CorpFin v2",
249
+ "benchmark_updated": "2026-06-17",
250
+ "task_key": "shared_max_context",
251
+ "task_name": "Shared Max Context",
252
+ "dataset_type": "private",
253
+ "industry": "finance",
254
+ "raw_score": "21.445",
255
+ "raw_stderr": "1.401",
256
+ "latency": "0.059",
257
+ "cost_per_test": "0.005006",
258
+ "max_output_tokens": "8192",
259
+ "provider": "Google"
260
+ },
261
+ "uncertainty": {
262
+ "standard_error": {
263
+ "value": 1.401,
264
+ "method": "vals_reported"
265
+ }
266
+ }
267
+ },
268
+ "generation_config": {
269
+ "generation_args": {
270
+ "max_tokens": 8192,
271
+ "max_attempts": 1
272
+ }
273
+ }
274
+ }
275
+ ]
276
+ }
data/vals_ai/google/gemini-1.5-flash-002/03b06b69-5f43-40ea-9cd2-21a044e0ca86.json ADDED
@@ -0,0 +1,981 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmlu_pro/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMLU Pro",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmlu_pro",
13
+ "benchmark_name": "MMLU Pro",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmlu_pro:biology:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.mmlu_pro.biology",
38
+ "source_data": {
39
+ "dataset_name": "MMLU Pro - Biology",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmlu_pro"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmlu_pro",
46
+ "task_key": "biology",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Biology).",
53
+ "metric_id": "vals_ai.mmlu_pro.biology.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 82.706,
69
+ "details": {
70
+ "benchmark_slug": "mmlu_pro",
71
+ "benchmark_name": "MMLU Pro",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "biology",
74
+ "task_name": "Biology",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "82.706",
78
+ "raw_stderr": "1.421",
79
+ "latency": "1.504",
80
+ "cost_per_test": "0.000164",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.421,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mmlu_pro:business:google/gemini-1.5-flash-002:score",
100
+ "evaluation_name": "vals_ai.mmlu_pro.business",
101
+ "source_data": {
102
+ "dataset_name": "MMLU Pro - Business",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mmlu_pro"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mmlu_pro",
109
+ "task_key": "business",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Business).",
116
+ "metric_id": "vals_ai.mmlu_pro.business.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 71.61,
132
+ "details": {
133
+ "benchmark_slug": "mmlu_pro",
134
+ "benchmark_name": "MMLU Pro",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "business",
137
+ "task_name": "Business",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "71.61",
141
+ "raw_stderr": "1.607",
142
+ "latency": "1.793",
143
+ "cost_per_test": "0.000151",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.607,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mmlu_pro:chemistry:google/gemini-1.5-flash-002:score",
163
+ "evaluation_name": "vals_ai.mmlu_pro.chemistry",
164
+ "source_data": {
165
+ "dataset_name": "MMLU Pro - Chemistry",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mmlu_pro"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mmlu_pro",
172
+ "task_key": "chemistry",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Chemistry).",
179
+ "metric_id": "vals_ai.mmlu_pro.chemistry.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 71.555,
195
+ "details": {
196
+ "benchmark_slug": "mmlu_pro",
197
+ "benchmark_name": "MMLU Pro",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "chemistry",
200
+ "task_name": "Chemistry",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "71.555",
204
+ "raw_stderr": "1.341",
205
+ "latency": "2.534",
206
+ "cost_per_test": "0.00023",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.341,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mmlu_pro:computer_science:google/gemini-1.5-flash-002:score",
226
+ "evaluation_name": "vals_ai.mmlu_pro.computer_science",
227
+ "source_data": {
228
+ "dataset_name": "MMLU Pro - Computer Science",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mmlu_pro"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mmlu_pro",
235
+ "task_key": "computer_science",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Computer Science).",
242
+ "metric_id": "vals_ai.mmlu_pro.computer_science.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 66.098,
258
+ "details": {
259
+ "benchmark_slug": "mmlu_pro",
260
+ "benchmark_name": "MMLU Pro",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "computer_science",
263
+ "task_name": "Computer Science",
264
+ "dataset_type": "public",
265
+ "industry": "academic",
266
+ "raw_score": "66.098",
267
+ "raw_stderr": "2.346",
268
+ "latency": "1.769",
269
+ "cost_per_test": "0.000191",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 2.346,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mmlu_pro:economics:google/gemini-1.5-flash-002:score",
289
+ "evaluation_name": "vals_ai.mmlu_pro.economics",
290
+ "source_data": {
291
+ "dataset_name": "MMLU Pro - Economics",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mmlu_pro"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mmlu_pro",
298
+ "task_key": "economics",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Economics).",
305
+ "metric_id": "vals_ai.mmlu_pro.economics.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 73.934,
321
+ "details": {
322
+ "benchmark_slug": "mmlu_pro",
323
+ "benchmark_name": "MMLU Pro",
324
+ "benchmark_updated": "2026-06-17",
325
+ "task_key": "economics",
326
+ "task_name": "Economics",
327
+ "dataset_type": "public",
328
+ "industry": "academic",
329
+ "raw_score": "73.934",
330
+ "raw_stderr": "1.513",
331
+ "latency": "1.376",
332
+ "cost_per_test": "0.000149",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 1.513,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mmlu_pro:engineering:google/gemini-1.5-flash-002:score",
352
+ "evaluation_name": "vals_ai.mmlu_pro.engineering",
353
+ "source_data": {
354
+ "dataset_name": "MMLU Pro - Engineering",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mmlu_pro"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mmlu_pro",
361
+ "task_key": "engineering",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Engineering).",
368
+ "metric_id": "vals_ai.mmlu_pro.engineering.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 44.788,
384
+ "details": {
385
+ "benchmark_slug": "mmlu_pro",
386
+ "benchmark_name": "MMLU Pro",
387
+ "benchmark_updated": "2026-06-17",
388
+ "task_key": "engineering",
389
+ "task_name": "Engineering",
390
+ "dataset_type": "public",
391
+ "industry": "academic",
392
+ "raw_score": "44.788",
393
+ "raw_stderr": "1.597",
394
+ "latency": "2.879",
395
+ "cost_per_test": "0.000227",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 1.597,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mmlu_pro:health:google/gemini-1.5-flash-002:score",
415
+ "evaluation_name": "vals_ai.mmlu_pro.health",
416
+ "source_data": {
417
+ "dataset_name": "MMLU Pro - Health",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mmlu_pro"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mmlu_pro",
424
+ "task_key": "health",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Health).",
431
+ "metric_id": "vals_ai.mmlu_pro.health.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 62.714,
447
+ "details": {
448
+ "benchmark_slug": "mmlu_pro",
449
+ "benchmark_name": "MMLU Pro",
450
+ "benchmark_updated": "2026-06-17",
451
+ "task_key": "health",
452
+ "task_name": "Health",
453
+ "dataset_type": "public",
454
+ "industry": "academic",
455
+ "raw_score": "62.714",
456
+ "raw_stderr": "1.691",
457
+ "latency": "1.396",
458
+ "cost_per_test": "0.000135",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.691,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mmlu_pro:history:google/gemini-1.5-flash-002:score",
478
+ "evaluation_name": "vals_ai.mmlu_pro.history",
479
+ "source_data": {
480
+ "dataset_name": "MMLU Pro - History",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mmlu_pro"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mmlu_pro",
487
+ "task_key": "history",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (History).",
494
+ "metric_id": "vals_ai.mmlu_pro.history.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 58.53,
510
+ "details": {
511
+ "benchmark_slug": "mmlu_pro",
512
+ "benchmark_name": "MMLU Pro",
513
+ "benchmark_updated": "2026-06-17",
514
+ "task_key": "history",
515
+ "task_name": "History",
516
+ "dataset_type": "public",
517
+ "industry": "academic",
518
+ "raw_score": "58.53",
519
+ "raw_stderr": "2.524",
520
+ "latency": "1.045",
521
+ "cost_per_test": "0.000199",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 2.524,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mmlu_pro:law:google/gemini-1.5-flash-002:score",
541
+ "evaluation_name": "vals_ai.mmlu_pro.law",
542
+ "source_data": {
543
+ "dataset_name": "MMLU Pro - Law",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mmlu_pro"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mmlu_pro",
550
+ "task_key": "law",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Law).",
557
+ "metric_id": "vals_ai.mmlu_pro.law.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 42.598,
573
+ "details": {
574
+ "benchmark_slug": "mmlu_pro",
575
+ "benchmark_name": "MMLU Pro",
576
+ "benchmark_updated": "2026-06-17",
577
+ "task_key": "law",
578
+ "task_name": "Law",
579
+ "dataset_type": "public",
580
+ "industry": "academic",
581
+ "raw_score": "42.598",
582
+ "raw_stderr": "1.49",
583
+ "latency": "1.323",
584
+ "cost_per_test": "0.000161",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 1.49,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mmlu_pro:math:google/gemini-1.5-flash-002:score",
604
+ "evaluation_name": "vals_ai.mmlu_pro.math",
605
+ "source_data": {
606
+ "dataset_name": "MMLU Pro - Math",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mmlu_pro"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mmlu_pro",
613
+ "task_key": "math",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Math).",
620
+ "metric_id": "vals_ai.mmlu_pro.math.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 79.201,
636
+ "details": {
637
+ "benchmark_slug": "mmlu_pro",
638
+ "benchmark_name": "MMLU Pro",
639
+ "benchmark_updated": "2026-06-17",
640
+ "task_key": "math",
641
+ "task_name": "Math",
642
+ "dataset_type": "public",
643
+ "industry": "academic",
644
+ "raw_score": "79.201",
645
+ "raw_stderr": "1.113",
646
+ "latency": "2.126",
647
+ "cost_per_test": "0.000223",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 1.113,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mmlu_pro:other:google/gemini-1.5-flash-002:score",
667
+ "evaluation_name": "vals_ai.mmlu_pro.other",
668
+ "source_data": {
669
+ "dataset_name": "MMLU Pro - Others",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mmlu_pro"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mmlu_pro",
676
+ "task_key": "other",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Others).",
683
+ "metric_id": "vals_ai.mmlu_pro.other.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 62.338,
699
+ "details": {
700
+ "benchmark_slug": "mmlu_pro",
701
+ "benchmark_name": "MMLU Pro",
702
+ "benchmark_updated": "2026-06-17",
703
+ "task_key": "other",
704
+ "task_name": "Others",
705
+ "dataset_type": "public",
706
+ "industry": "academic",
707
+ "raw_score": "62.338",
708
+ "raw_stderr": "1.594",
709
+ "latency": "1.045",
710
+ "cost_per_test": "0.000108",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 1.594,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mmlu_pro:overall:google/gemini-1.5-flash-002:score",
730
+ "evaluation_name": "vals_ai.mmlu_pro.overall",
731
+ "source_data": {
732
+ "dataset_name": "MMLU Pro - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mmlu_pro"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mmlu_pro",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Overall).",
746
+ "metric_id": "vals_ai.mmlu_pro.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 65.606,
762
+ "details": {
763
+ "benchmark_slug": "mmlu_pro",
764
+ "benchmark_name": "MMLU Pro",
765
+ "benchmark_updated": "2026-06-17",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "academic",
770
+ "raw_score": "65.606",
771
+ "raw_stderr": "0.456",
772
+ "latency": "1.682",
773
+ "cost_per_test": "0.00017",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 0.456,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ },
791
+ {
792
+ "evaluation_result_id": "mmlu_pro:philosophy:google/gemini-1.5-flash-002:score",
793
+ "evaluation_name": "vals_ai.mmlu_pro.philosophy",
794
+ "source_data": {
795
+ "dataset_name": "MMLU Pro - Philosophy",
796
+ "source_type": "url",
797
+ "url": [
798
+ "https://www.vals.ai/benchmarks/mmlu_pro"
799
+ ],
800
+ "additional_details": {
801
+ "benchmark_slug": "mmlu_pro",
802
+ "task_key": "philosophy",
803
+ "dataset_type": "public",
804
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
805
+ }
806
+ },
807
+ "metric_config": {
808
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Philosophy).",
809
+ "metric_id": "vals_ai.mmlu_pro.philosophy.accuracy",
810
+ "metric_name": "Accuracy",
811
+ "metric_kind": "accuracy",
812
+ "metric_unit": "percent",
813
+ "lower_is_better": false,
814
+ "score_type": "continuous",
815
+ "min_score": 0.0,
816
+ "max_score": 100.0,
817
+ "additional_details": {
818
+ "score_scale": "percent_0_to_100",
819
+ "max_score_source": "fixed_percentage_bound",
820
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
821
+ }
822
+ },
823
+ "score_details": {
824
+ "score": 56.513,
825
+ "details": {
826
+ "benchmark_slug": "mmlu_pro",
827
+ "benchmark_name": "MMLU Pro",
828
+ "benchmark_updated": "2026-06-17",
829
+ "task_key": "philosophy",
830
+ "task_name": "Philosophy",
831
+ "dataset_type": "public",
832
+ "industry": "academic",
833
+ "raw_score": "56.513",
834
+ "raw_stderr": "2.219",
835
+ "latency": "1.282",
836
+ "cost_per_test": "0.000123",
837
+ "max_output_tokens": "8192",
838
+ "provider": "Google"
839
+ },
840
+ "uncertainty": {
841
+ "standard_error": {
842
+ "value": 2.219,
843
+ "method": "vals_reported"
844
+ }
845
+ }
846
+ },
847
+ "generation_config": {
848
+ "generation_args": {
849
+ "max_tokens": 8192,
850
+ "max_attempts": 1
851
+ }
852
+ }
853
+ },
854
+ {
855
+ "evaluation_result_id": "mmlu_pro:physics:google/gemini-1.5-flash-002:score",
856
+ "evaluation_name": "vals_ai.mmlu_pro.physics",
857
+ "source_data": {
858
+ "dataset_name": "MMLU Pro - Physics",
859
+ "source_type": "url",
860
+ "url": [
861
+ "https://www.vals.ai/benchmarks/mmlu_pro"
862
+ ],
863
+ "additional_details": {
864
+ "benchmark_slug": "mmlu_pro",
865
+ "task_key": "physics",
866
+ "dataset_type": "public",
867
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
868
+ }
869
+ },
870
+ "metric_config": {
871
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Physics).",
872
+ "metric_id": "vals_ai.mmlu_pro.physics.accuracy",
873
+ "metric_name": "Accuracy",
874
+ "metric_kind": "accuracy",
875
+ "metric_unit": "percent",
876
+ "lower_is_better": false,
877
+ "score_type": "continuous",
878
+ "min_score": 0.0,
879
+ "max_score": 100.0,
880
+ "additional_details": {
881
+ "score_scale": "percent_0_to_100",
882
+ "max_score_source": "fixed_percentage_bound",
883
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
884
+ }
885
+ },
886
+ "score_details": {
887
+ "score": 71.209,
888
+ "details": {
889
+ "benchmark_slug": "mmlu_pro",
890
+ "benchmark_name": "MMLU Pro",
891
+ "benchmark_updated": "2026-06-17",
892
+ "task_key": "physics",
893
+ "task_name": "Physics",
894
+ "dataset_type": "public",
895
+ "industry": "academic",
896
+ "raw_score": "71.209",
897
+ "raw_stderr": "1.256",
898
+ "latency": "2.297",
899
+ "cost_per_test": "0.000188",
900
+ "max_output_tokens": "8192",
901
+ "provider": "Google"
902
+ },
903
+ "uncertainty": {
904
+ "standard_error": {
905
+ "value": 1.256,
906
+ "method": "vals_reported"
907
+ }
908
+ }
909
+ },
910
+ "generation_config": {
911
+ "generation_args": {
912
+ "max_tokens": 8192,
913
+ "max_attempts": 1
914
+ }
915
+ }
916
+ },
917
+ {
918
+ "evaluation_result_id": "mmlu_pro:psychology:google/gemini-1.5-flash-002:score",
919
+ "evaluation_name": "vals_ai.mmlu_pro.psychology",
920
+ "source_data": {
921
+ "dataset_name": "MMLU Pro - Psychology",
922
+ "source_type": "url",
923
+ "url": [
924
+ "https://www.vals.ai/benchmarks/mmlu_pro"
925
+ ],
926
+ "additional_details": {
927
+ "benchmark_slug": "mmlu_pro",
928
+ "task_key": "psychology",
929
+ "dataset_type": "public",
930
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
931
+ }
932
+ },
933
+ "metric_config": {
934
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Psychology).",
935
+ "metric_id": "vals_ai.mmlu_pro.psychology.accuracy",
936
+ "metric_name": "Accuracy",
937
+ "metric_kind": "accuracy",
938
+ "metric_unit": "percent",
939
+ "lower_is_better": false,
940
+ "score_type": "continuous",
941
+ "min_score": 0.0,
942
+ "max_score": 100.0,
943
+ "additional_details": {
944
+ "score_scale": "percent_0_to_100",
945
+ "max_score_source": "fixed_percentage_bound",
946
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
947
+ }
948
+ },
949
+ "score_details": {
950
+ "score": 74.687,
951
+ "details": {
952
+ "benchmark_slug": "mmlu_pro",
953
+ "benchmark_name": "MMLU Pro",
954
+ "benchmark_updated": "2026-06-17",
955
+ "task_key": "psychology",
956
+ "task_name": "Psychology",
957
+ "dataset_type": "public",
958
+ "industry": "academic",
959
+ "raw_score": "74.687",
960
+ "raw_stderr": "1.539",
961
+ "latency": "1.181",
962
+ "cost_per_test": "0.000126",
963
+ "max_output_tokens": "8192",
964
+ "provider": "Google"
965
+ },
966
+ "uncertainty": {
967
+ "standard_error": {
968
+ "value": 1.539,
969
+ "method": "vals_reported"
970
+ }
971
+ }
972
+ },
973
+ "generation_config": {
974
+ "generation_args": {
975
+ "max_tokens": 8192,
976
+ "max_attempts": 1
977
+ }
978
+ }
979
+ }
980
+ ]
981
+ }
data/vals_ai/google/gemini-1.5-flash-002/15f5e37d-a115-4d47-ad22-5d511fd7cff7.json ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/gpqa/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - GPQA Diamond",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "gpqa",
13
+ "benchmark_name": "GPQA Diamond",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "gpqa:diamond_few_shot_cot:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.gpqa.diamond_few_shot_cot",
38
+ "source_data": {
39
+ "dataset_name": "GPQA Diamond - Few-Shot CoT",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/gpqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "gpqa",
46
+ "task_key": "diamond_few_shot_cot",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Few-Shot CoT).",
53
+ "metric_id": "vals_ai.gpqa.diamond_few_shot_cot.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 44.949,
69
+ "details": {
70
+ "benchmark_slug": "gpqa",
71
+ "benchmark_name": "GPQA Diamond",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "diamond_few_shot_cot",
74
+ "task_name": "Few-Shot CoT",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "44.949",
78
+ "raw_stderr": "3.535",
79
+ "latency": "3.47",
80
+ "cost_per_test": "0.000299",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.535,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "gpqa:diamond_zero_shot_cot:google/gemini-1.5-flash-002:score",
100
+ "evaluation_name": "vals_ai.gpqa.diamond_zero_shot_cot",
101
+ "source_data": {
102
+ "dataset_name": "GPQA Diamond - Zero-Shot CoT",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/gpqa"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "gpqa",
109
+ "task_key": "diamond_zero_shot_cot",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Zero-Shot CoT).",
116
+ "metric_id": "vals_ai.gpqa.diamond_zero_shot_cot.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 46.97,
132
+ "details": {
133
+ "benchmark_slug": "gpqa",
134
+ "benchmark_name": "GPQA Diamond",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "diamond_zero_shot_cot",
137
+ "task_name": "Zero-Shot CoT",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "46.97",
141
+ "raw_stderr": "3.547",
142
+ "latency": "3.311",
143
+ "cost_per_test": "0.000167",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 3.547,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "gpqa:overall:google/gemini-1.5-flash-002:score",
163
+ "evaluation_name": "vals_ai.gpqa.overall",
164
+ "source_data": {
165
+ "dataset_name": "GPQA Diamond - Overall",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/gpqa"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "gpqa",
172
+ "task_key": "overall",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Overall).",
179
+ "metric_id": "vals_ai.gpqa.overall.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 45.96,
195
+ "details": {
196
+ "benchmark_slug": "gpqa",
197
+ "benchmark_name": "GPQA Diamond",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "overall",
200
+ "task_name": "Overall",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "45.96",
204
+ "raw_stderr": "2.504",
205
+ "latency": "3.39",
206
+ "cost_per_test": "0.000233",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.504,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ }
224
+ ]
225
+ }
data/vals_ai/google/gemini-1.5-flash-002/2850f115-e5ce-42ce-a5bf-8bcaf55a091a.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 36.141,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "36.141",
75
+ "raw_stderr": "1.374",
76
+ "latency": "3.701",
77
+ "cost_per_test": "0.000155",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.374,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-1.5-flash-002:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 48.202,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "48.202",
135
+ "raw_stderr": "0.98",
136
+ "latency": "3.701",
137
+ "cost_per_test": "0.000155",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.98,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-1.5-flash-002:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 60.262,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "60.262",
195
+ "raw_stderr": "1.399",
196
+ "latency": "3.701",
197
+ "cost_per_test": "0.000155",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.399,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-1.5-flash-002/4ee05046-cf52-4301-a477-282ede77be12.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/aime/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - AIME",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "aime",
13
+ "benchmark_name": "AIME",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "aime:aime_2024:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.aime.aime_2024",
38
+ "source_data": {
39
+ "dataset_name": "AIME - AIME 2024",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/aime"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "aime",
46
+ "task_key": "aime_2024",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2024).",
53
+ "metric_id": "vals_ai.aime.aime_2024.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 17.083,
69
+ "details": {
70
+ "benchmark_slug": "aime",
71
+ "benchmark_name": "AIME",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "aime_2024",
74
+ "task_name": "AIME 2024",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "17.083",
78
+ "raw_stderr": "1.327",
79
+ "latency": "5.826",
80
+ "cost_per_test": "0.000271",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 1.327,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ }
90
+ },
91
+ {
92
+ "evaluation_result_id": "aime:aime_2025:google/gemini-1.5-flash-002:score",
93
+ "evaluation_name": "vals_ai.aime.aime_2025",
94
+ "source_data": {
95
+ "dataset_name": "AIME - AIME 2025",
96
+ "source_type": "url",
97
+ "url": [
98
+ "https://www.vals.ai/benchmarks/aime"
99
+ ],
100
+ "additional_details": {
101
+ "benchmark_slug": "aime",
102
+ "task_key": "aime_2025",
103
+ "dataset_type": "public",
104
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
105
+ }
106
+ },
107
+ "metric_config": {
108
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2025).",
109
+ "metric_id": "vals_ai.aime.aime_2025.accuracy",
110
+ "metric_name": "Accuracy",
111
+ "metric_kind": "accuracy",
112
+ "metric_unit": "percent",
113
+ "lower_is_better": false,
114
+ "score_type": "continuous",
115
+ "min_score": 0.0,
116
+ "max_score": 100.0,
117
+ "additional_details": {
118
+ "score_scale": "percent_0_to_100",
119
+ "max_score_source": "fixed_percentage_bound",
120
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
121
+ }
122
+ },
123
+ "score_details": {
124
+ "score": 17.5,
125
+ "details": {
126
+ "benchmark_slug": "aime",
127
+ "benchmark_name": "AIME",
128
+ "benchmark_updated": "2026-04-16",
129
+ "task_key": "aime_2025",
130
+ "task_name": "AIME 2025",
131
+ "dataset_type": "public",
132
+ "industry": "math",
133
+ "raw_score": "17.5",
134
+ "raw_stderr": "1.22",
135
+ "latency": "5.567",
136
+ "cost_per_test": "0.00026",
137
+ "provider": "Google"
138
+ },
139
+ "uncertainty": {
140
+ "standard_error": {
141
+ "value": 1.22,
142
+ "method": "vals_reported"
143
+ }
144
+ }
145
+ }
146
+ },
147
+ {
148
+ "evaluation_result_id": "aime:overall:google/gemini-1.5-flash-002:score",
149
+ "evaluation_name": "vals_ai.aime.overall",
150
+ "source_data": {
151
+ "dataset_name": "AIME - Overall",
152
+ "source_type": "url",
153
+ "url": [
154
+ "https://www.vals.ai/benchmarks/aime"
155
+ ],
156
+ "additional_details": {
157
+ "benchmark_slug": "aime",
158
+ "task_key": "overall",
159
+ "dataset_type": "public",
160
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
161
+ }
162
+ },
163
+ "metric_config": {
164
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (Overall).",
165
+ "metric_id": "vals_ai.aime.overall.accuracy",
166
+ "metric_name": "Accuracy",
167
+ "metric_kind": "accuracy",
168
+ "metric_unit": "percent",
169
+ "lower_is_better": false,
170
+ "score_type": "continuous",
171
+ "min_score": 0.0,
172
+ "max_score": 100.0,
173
+ "additional_details": {
174
+ "score_scale": "percent_0_to_100",
175
+ "max_score_source": "fixed_percentage_bound",
176
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
177
+ }
178
+ },
179
+ "score_details": {
180
+ "score": 17.292,
181
+ "details": {
182
+ "benchmark_slug": "aime",
183
+ "benchmark_name": "AIME",
184
+ "benchmark_updated": "2026-04-16",
185
+ "task_key": "overall",
186
+ "task_name": "Overall",
187
+ "dataset_type": "public",
188
+ "industry": "math",
189
+ "raw_score": "17.292",
190
+ "raw_stderr": "0.901",
191
+ "latency": "5.696",
192
+ "cost_per_test": "0.000266",
193
+ "provider": "Google"
194
+ },
195
+ "uncertainty": {
196
+ "standard_error": {
197
+ "value": 0.901,
198
+ "method": "vals_reported"
199
+ }
200
+ }
201
+ }
202
+ }
203
+ ]
204
+ }
data/vals_ai/google/gemini-1.5-flash-002/6672691c-ff0e-409f-a8b4-ef6ac0316ce9.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mortgage_tax/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MortgageTax",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mortgage_tax",
13
+ "benchmark_name": "MortgageTax",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mortgage_tax:numerical_extraction:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.mortgage_tax.numerical_extraction",
38
+ "source_data": {
39
+ "dataset_name": "MortgageTax - Numerical Extraction",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "mortgage_tax",
43
+ "task_key": "numerical_extraction",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Numerical Extraction).",
50
+ "metric_id": "vals_ai.mortgage_tax.numerical_extraction.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 58.108,
66
+ "details": {
67
+ "benchmark_slug": "mortgage_tax",
68
+ "benchmark_name": "MortgageTax",
69
+ "benchmark_updated": "2026-06-10",
70
+ "task_key": "numerical_extraction",
71
+ "task_name": "Numerical Extraction",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "58.108",
75
+ "raw_stderr": "2.722",
76
+ "latency": "2.002",
77
+ "cost_per_test": "0.00011",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 2.722,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "mortgage_tax:overall:google/gemini-1.5-flash-002:score",
97
+ "evaluation_name": "vals_ai.mortgage_tax.overall",
98
+ "source_data": {
99
+ "dataset_name": "MortgageTax - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "mortgage_tax",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Overall).",
110
+ "metric_id": "vals_ai.mortgage_tax.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 42.766,
126
+ "details": {
127
+ "benchmark_slug": "mortgage_tax",
128
+ "benchmark_name": "MortgageTax",
129
+ "benchmark_updated": "2026-06-10",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "42.766",
135
+ "raw_stderr": "1.835",
136
+ "latency": "2.1",
137
+ "cost_per_test": "0.00012",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.835,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "mortgage_tax:semantic_extraction:google/gemini-1.5-flash-002:score",
157
+ "evaluation_name": "vals_ai.mortgage_tax.semantic_extraction",
158
+ "source_data": {
159
+ "dataset_name": "MortgageTax - Semantic Extraction",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "mortgage_tax",
163
+ "task_key": "semantic_extraction",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Semantic Extraction).",
170
+ "metric_id": "vals_ai.mortgage_tax.semantic_extraction.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 27.424,
186
+ "details": {
187
+ "benchmark_slug": "mortgage_tax",
188
+ "benchmark_name": "MortgageTax",
189
+ "benchmark_updated": "2026-06-10",
190
+ "task_key": "semantic_extraction",
191
+ "task_name": "Semantic Extraction",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "27.424",
195
+ "raw_stderr": "2.463",
196
+ "latency": "2.197",
197
+ "cost_per_test": "0.000129",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 2.463,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-1.5-flash-002/7e409542-e6e4-4bd0-87bc-2fbfe8e1f846.json ADDED
@@ -0,0 +1,408 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 78.095,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "78.095",
78
+ "raw_stderr": "1.028",
79
+ "latency": "0.36",
80
+ "max_output_tokens": "8192",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 1.028,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 8192,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "legal_bench:interpretation_tasks:google/gemini-1.5-flash-002:score",
99
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
100
+ "source_data": {
101
+ "dataset_name": "LegalBench - Interpretation Tasks",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/legal_bench"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "legal_bench",
108
+ "task_key": "interpretation_tasks",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
115
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 61.847,
131
+ "details": {
132
+ "benchmark_slug": "legal_bench",
133
+ "benchmark_name": "LegalBench",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "interpretation_tasks",
136
+ "task_name": "Interpretation Tasks",
137
+ "dataset_type": "public",
138
+ "industry": "legal",
139
+ "raw_score": "61.847",
140
+ "raw_stderr": "0.208",
141
+ "latency": "0.404",
142
+ "max_output_tokens": "8192",
143
+ "provider": "Google"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 0.208,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 8192,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "legal_bench:issue_tasks:google/gemini-1.5-flash-002:score",
161
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
162
+ "source_data": {
163
+ "dataset_name": "LegalBench - Issue Tasks",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/legal_bench"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "legal_bench",
170
+ "task_key": "issue_tasks",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
177
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 51.973,
193
+ "details": {
194
+ "benchmark_slug": "legal_bench",
195
+ "benchmark_name": "LegalBench",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "issue_tasks",
198
+ "task_name": "Issue Tasks",
199
+ "dataset_type": "public",
200
+ "industry": "legal",
201
+ "raw_score": "51.973",
202
+ "raw_stderr": "0.263",
203
+ "latency": "0.445",
204
+ "max_output_tokens": "8192",
205
+ "provider": "Google"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 0.263,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 8192,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "legal_bench:overall:google/gemini-1.5-flash-002:score",
223
+ "evaluation_name": "vals_ai.legal_bench.overall",
224
+ "source_data": {
225
+ "dataset_name": "LegalBench - Overall",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/legal_bench"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "legal_bench",
232
+ "task_key": "overall",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
239
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 63.238,
255
+ "details": {
256
+ "benchmark_slug": "legal_bench",
257
+ "benchmark_name": "LegalBench",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": "overall",
260
+ "task_name": "Overall",
261
+ "dataset_type": "public",
262
+ "industry": "legal",
263
+ "raw_score": "63.238",
264
+ "raw_stderr": "0.736",
265
+ "latency": "0.476",
266
+ "max_output_tokens": "8192",
267
+ "provider": "Google"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 0.736,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 8192,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ },
283
+ {
284
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:google/gemini-1.5-flash-002:score",
285
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
286
+ "source_data": {
287
+ "dataset_name": "LegalBench - Rhetoric Tasks",
288
+ "source_type": "url",
289
+ "url": [
290
+ "https://www.vals.ai/benchmarks/legal_bench"
291
+ ],
292
+ "additional_details": {
293
+ "benchmark_slug": "legal_bench",
294
+ "task_key": "rhetoric_tasks",
295
+ "dataset_type": "public",
296
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
297
+ }
298
+ },
299
+ "metric_config": {
300
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
301
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
302
+ "metric_name": "Accuracy",
303
+ "metric_kind": "accuracy",
304
+ "metric_unit": "percent",
305
+ "lower_is_better": false,
306
+ "score_type": "continuous",
307
+ "min_score": 0.0,
308
+ "max_score": 100.0,
309
+ "additional_details": {
310
+ "score_scale": "percent_0_to_100",
311
+ "max_score_source": "fixed_percentage_bound",
312
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
313
+ }
314
+ },
315
+ "score_details": {
316
+ "score": 61.657,
317
+ "details": {
318
+ "benchmark_slug": "legal_bench",
319
+ "benchmark_name": "LegalBench",
320
+ "benchmark_updated": "2026-06-17",
321
+ "task_key": "rhetoric_tasks",
322
+ "task_name": "Rhetoric Tasks",
323
+ "dataset_type": "public",
324
+ "industry": "legal",
325
+ "raw_score": "61.657",
326
+ "raw_stderr": "0.667",
327
+ "latency": "0.397",
328
+ "max_output_tokens": "8192",
329
+ "provider": "Google"
330
+ },
331
+ "uncertainty": {
332
+ "standard_error": {
333
+ "value": 0.667,
334
+ "method": "vals_reported"
335
+ }
336
+ }
337
+ },
338
+ "generation_config": {
339
+ "generation_args": {
340
+ "max_tokens": 8192,
341
+ "max_attempts": 1
342
+ }
343
+ }
344
+ },
345
+ {
346
+ "evaluation_result_id": "legal_bench:rule_tasks:google/gemini-1.5-flash-002:score",
347
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
348
+ "source_data": {
349
+ "dataset_name": "LegalBench - Rule Tasks",
350
+ "source_type": "url",
351
+ "url": [
352
+ "https://www.vals.ai/benchmarks/legal_bench"
353
+ ],
354
+ "additional_details": {
355
+ "benchmark_slug": "legal_bench",
356
+ "task_key": "rule_tasks",
357
+ "dataset_type": "public",
358
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
359
+ }
360
+ },
361
+ "metric_config": {
362
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
363
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
364
+ "metric_name": "Accuracy",
365
+ "metric_kind": "accuracy",
366
+ "metric_unit": "percent",
367
+ "lower_is_better": false,
368
+ "score_type": "continuous",
369
+ "min_score": 0.0,
370
+ "max_score": 100.0,
371
+ "additional_details": {
372
+ "score_scale": "percent_0_to_100",
373
+ "max_score_source": "fixed_percentage_bound",
374
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
375
+ }
376
+ },
377
+ "score_details": {
378
+ "score": 62.618,
379
+ "details": {
380
+ "benchmark_slug": "legal_bench",
381
+ "benchmark_name": "LegalBench",
382
+ "benchmark_updated": "2026-06-17",
383
+ "task_key": "rule_tasks",
384
+ "task_name": "Rule Tasks",
385
+ "dataset_type": "public",
386
+ "industry": "legal",
387
+ "raw_score": "62.618",
388
+ "raw_stderr": "3.456",
389
+ "latency": "0.772",
390
+ "max_output_tokens": "8192",
391
+ "provider": "Google"
392
+ },
393
+ "uncertainty": {
394
+ "standard_error": {
395
+ "value": 3.456,
396
+ "method": "vals_reported"
397
+ }
398
+ }
399
+ },
400
+ "generation_config": {
401
+ "generation_args": {
402
+ "max_tokens": 8192,
403
+ "max_attempts": 1
404
+ }
405
+ }
406
+ }
407
+ ]
408
+ }
data/vals_ai/google/gemini-1.5-flash-002/84088ee9-a4a6-4cf2-8d1d-4711ab9ed15a.json ADDED
@@ -0,0 +1,792 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mgsm/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MGSM",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mgsm",
13
+ "benchmark_name": "MGSM",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mgsm:mgsm_bn:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.mgsm.mgsm_bn",
38
+ "source_data": {
39
+ "dataset_name": "MGSM - Bengali",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mgsm"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mgsm",
46
+ "task_key": "mgsm_bn",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Bengali).",
53
+ "metric_id": "vals_ai.mgsm.mgsm_bn.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 85.6,
69
+ "details": {
70
+ "benchmark_slug": "mgsm",
71
+ "benchmark_name": "MGSM",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "mgsm_bn",
74
+ "task_name": "Bengali",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "85.6",
78
+ "raw_stderr": "2.22",
79
+ "latency": "1.932",
80
+ "cost_per_test": "9.5e-05",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 2.22,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mgsm:mgsm_de:google/gemini-1.5-flash-002:score",
100
+ "evaluation_name": "vals_ai.mgsm.mgsm_de",
101
+ "source_data": {
102
+ "dataset_name": "MGSM - German",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mgsm"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mgsm",
109
+ "task_key": "mgsm_de",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (German).",
116
+ "metric_id": "vals_ai.mgsm.mgsm_de.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 88.4,
132
+ "details": {
133
+ "benchmark_slug": "mgsm",
134
+ "benchmark_name": "MGSM",
135
+ "benchmark_updated": "2026-01-09",
136
+ "task_key": "mgsm_de",
137
+ "task_name": "German",
138
+ "dataset_type": "public",
139
+ "industry": "math",
140
+ "raw_score": "88.4",
141
+ "raw_stderr": "2.025",
142
+ "latency": "1.511",
143
+ "cost_per_test": "6.5e-05",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 2.025,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mgsm:mgsm_en:google/gemini-1.5-flash-002:score",
163
+ "evaluation_name": "vals_ai.mgsm.mgsm_en",
164
+ "source_data": {
165
+ "dataset_name": "MGSM - English",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mgsm"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mgsm",
172
+ "task_key": "mgsm_en",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (English).",
179
+ "metric_id": "vals_ai.mgsm.mgsm_en.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 94.0,
195
+ "details": {
196
+ "benchmark_slug": "mgsm",
197
+ "benchmark_name": "MGSM",
198
+ "benchmark_updated": "2026-01-09",
199
+ "task_key": "mgsm_en",
200
+ "task_name": "English",
201
+ "dataset_type": "public",
202
+ "industry": "math",
203
+ "raw_score": "94",
204
+ "raw_stderr": "1.502",
205
+ "latency": "1.09",
206
+ "cost_per_test": "4.6e-05",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.502,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mgsm:mgsm_es:google/gemini-1.5-flash-002:score",
226
+ "evaluation_name": "vals_ai.mgsm.mgsm_es",
227
+ "source_data": {
228
+ "dataset_name": "MGSM - Spanish",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mgsm"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mgsm",
235
+ "task_key": "mgsm_es",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Spanish).",
242
+ "metric_id": "vals_ai.mgsm.mgsm_es.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 86.4,
258
+ "details": {
259
+ "benchmark_slug": "mgsm",
260
+ "benchmark_name": "MGSM",
261
+ "benchmark_updated": "2026-01-09",
262
+ "task_key": "mgsm_es",
263
+ "task_name": "Spanish",
264
+ "dataset_type": "public",
265
+ "industry": "math",
266
+ "raw_score": "86.4",
267
+ "raw_stderr": "2.168",
268
+ "latency": "1.127",
269
+ "cost_per_test": "4.8e-05",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 2.168,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mgsm:mgsm_fr:google/gemini-1.5-flash-002:score",
289
+ "evaluation_name": "vals_ai.mgsm.mgsm_fr",
290
+ "source_data": {
291
+ "dataset_name": "MGSM - French",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mgsm"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mgsm",
298
+ "task_key": "mgsm_fr",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (French).",
305
+ "metric_id": "vals_ai.mgsm.mgsm_fr.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 78.8,
321
+ "details": {
322
+ "benchmark_slug": "mgsm",
323
+ "benchmark_name": "MGSM",
324
+ "benchmark_updated": "2026-01-09",
325
+ "task_key": "mgsm_fr",
326
+ "task_name": "French",
327
+ "dataset_type": "public",
328
+ "industry": "math",
329
+ "raw_score": "78.8",
330
+ "raw_stderr": "2.585",
331
+ "latency": "1.173",
332
+ "cost_per_test": "5.2e-05",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 2.585,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mgsm:mgsm_ja:google/gemini-1.5-flash-002:score",
352
+ "evaluation_name": "vals_ai.mgsm.mgsm_ja",
353
+ "source_data": {
354
+ "dataset_name": "MGSM - Japanese",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mgsm"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mgsm",
361
+ "task_key": "mgsm_ja",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Japanese).",
368
+ "metric_id": "vals_ai.mgsm.mgsm_ja.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 84.4,
384
+ "details": {
385
+ "benchmark_slug": "mgsm",
386
+ "benchmark_name": "MGSM",
387
+ "benchmark_updated": "2026-01-09",
388
+ "task_key": "mgsm_ja",
389
+ "task_name": "Japanese",
390
+ "dataset_type": "public",
391
+ "industry": "math",
392
+ "raw_score": "84.4",
393
+ "raw_stderr": "2.295",
394
+ "latency": "1.288",
395
+ "cost_per_test": "5.6e-05",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 2.295,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mgsm:mgsm_ru:google/gemini-1.5-flash-002:score",
415
+ "evaluation_name": "vals_ai.mgsm.mgsm_ru",
416
+ "source_data": {
417
+ "dataset_name": "MGSM - Russian",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mgsm"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mgsm",
424
+ "task_key": "mgsm_ru",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Russian).",
431
+ "metric_id": "vals_ai.mgsm.mgsm_ru.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 90.4,
447
+ "details": {
448
+ "benchmark_slug": "mgsm",
449
+ "benchmark_name": "MGSM",
450
+ "benchmark_updated": "2026-01-09",
451
+ "task_key": "mgsm_ru",
452
+ "task_name": "Russian",
453
+ "dataset_type": "public",
454
+ "industry": "math",
455
+ "raw_score": "90.4",
456
+ "raw_stderr": "1.863",
457
+ "latency": "1.199",
458
+ "cost_per_test": "5.2e-05",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.863,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mgsm:mgsm_sw:google/gemini-1.5-flash-002:score",
478
+ "evaluation_name": "vals_ai.mgsm.mgsm_sw",
479
+ "source_data": {
480
+ "dataset_name": "MGSM - Swahili",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mgsm"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mgsm",
487
+ "task_key": "mgsm_sw",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Swahili).",
494
+ "metric_id": "vals_ai.mgsm.mgsm_sw.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 85.6,
510
+ "details": {
511
+ "benchmark_slug": "mgsm",
512
+ "benchmark_name": "MGSM",
513
+ "benchmark_updated": "2026-01-09",
514
+ "task_key": "mgsm_sw",
515
+ "task_name": "Swahili",
516
+ "dataset_type": "public",
517
+ "industry": "math",
518
+ "raw_score": "85.6",
519
+ "raw_stderr": "2.22",
520
+ "latency": "1.208",
521
+ "cost_per_test": "5.6e-05",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 2.22,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mgsm:mgsm_te:google/gemini-1.5-flash-002:score",
541
+ "evaluation_name": "vals_ai.mgsm.mgsm_te",
542
+ "source_data": {
543
+ "dataset_name": "MGSM - Telugu",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mgsm"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mgsm",
550
+ "task_key": "mgsm_te",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Telugu).",
557
+ "metric_id": "vals_ai.mgsm.mgsm_te.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 83.2,
573
+ "details": {
574
+ "benchmark_slug": "mgsm",
575
+ "benchmark_name": "MGSM",
576
+ "benchmark_updated": "2026-01-09",
577
+ "task_key": "mgsm_te",
578
+ "task_name": "Telugu",
579
+ "dataset_type": "public",
580
+ "industry": "math",
581
+ "raw_score": "83.2",
582
+ "raw_stderr": "2.364",
583
+ "latency": "2.49",
584
+ "cost_per_test": "0.000121",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 2.364,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mgsm:mgsm_th:google/gemini-1.5-flash-002:score",
604
+ "evaluation_name": "vals_ai.mgsm.mgsm_th",
605
+ "source_data": {
606
+ "dataset_name": "MGSM - Thai",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mgsm"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mgsm",
613
+ "task_key": "mgsm_th",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Thai).",
620
+ "metric_id": "vals_ai.mgsm.mgsm_th.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 88.0,
636
+ "details": {
637
+ "benchmark_slug": "mgsm",
638
+ "benchmark_name": "MGSM",
639
+ "benchmark_updated": "2026-01-09",
640
+ "task_key": "mgsm_th",
641
+ "task_name": "Thai",
642
+ "dataset_type": "public",
643
+ "industry": "math",
644
+ "raw_score": "88",
645
+ "raw_stderr": "2.055",
646
+ "latency": "1.289",
647
+ "cost_per_test": "5.8e-05",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 2.055,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mgsm:mgsm_zh:google/gemini-1.5-flash-002:score",
667
+ "evaluation_name": "vals_ai.mgsm.mgsm_zh",
668
+ "source_data": {
669
+ "dataset_name": "MGSM - Chinese",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mgsm"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mgsm",
676
+ "task_key": "mgsm_zh",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Chinese).",
683
+ "metric_id": "vals_ai.mgsm.mgsm_zh.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 87.6,
699
+ "details": {
700
+ "benchmark_slug": "mgsm",
701
+ "benchmark_name": "MGSM",
702
+ "benchmark_updated": "2026-01-09",
703
+ "task_key": "mgsm_zh",
704
+ "task_name": "Chinese",
705
+ "dataset_type": "public",
706
+ "industry": "math",
707
+ "raw_score": "87.6",
708
+ "raw_stderr": "2.084",
709
+ "latency": "1.233",
710
+ "cost_per_test": "5.2e-05",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 2.084,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mgsm:overall:google/gemini-1.5-flash-002:score",
730
+ "evaluation_name": "vals_ai.mgsm.overall",
731
+ "source_data": {
732
+ "dataset_name": "MGSM - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mgsm"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mgsm",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Overall).",
746
+ "metric_id": "vals_ai.mgsm.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 86.582,
762
+ "details": {
763
+ "benchmark_slug": "mgsm",
764
+ "benchmark_name": "MGSM",
765
+ "benchmark_updated": "2026-01-09",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "math",
770
+ "raw_score": "86.582",
771
+ "raw_stderr": "2.126",
772
+ "latency": "1.413",
773
+ "cost_per_test": "6.4e-05",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 2.126,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ }
791
+ ]
792
+ }
data/vals_ai/google/gemini-1.5-flash-002/8f75c4eb-0ceb-438e-89e2-fbac1d1b0424.json ADDED
@@ -0,0 +1,288 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/lcb/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LiveCodeBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "lcb",
13
+ "benchmark_name": "LiveCodeBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "lcb:easy:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.lcb.easy",
38
+ "source_data": {
39
+ "dataset_name": "LiveCodeBench - Easy",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/lcb"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "lcb",
46
+ "task_key": "easy",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Easy).",
53
+ "metric_id": "vals_ai.lcb.easy.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 82.609,
69
+ "details": {
70
+ "benchmark_slug": "lcb",
71
+ "benchmark_name": "LiveCodeBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "easy",
74
+ "task_name": "Easy",
75
+ "dataset_type": "public",
76
+ "industry": "coding",
77
+ "raw_score": "82.609",
78
+ "raw_stderr": "2.112",
79
+ "latency": "1.061",
80
+ "cost_per_test": "7.3e-05",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 2.112,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "lcb:hard:google/gemini-1.5-flash-002:score",
100
+ "evaluation_name": "vals_ai.lcb.hard",
101
+ "source_data": {
102
+ "dataset_name": "LiveCodeBench - Hard",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/lcb"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "lcb",
109
+ "task_key": "hard",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Hard).",
116
+ "metric_id": "vals_ai.lcb.hard.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 5.143,
132
+ "details": {
133
+ "benchmark_slug": "lcb",
134
+ "benchmark_name": "LiveCodeBench",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "hard",
137
+ "task_name": "Hard",
138
+ "dataset_type": "public",
139
+ "industry": "coding",
140
+ "raw_score": "5.143",
141
+ "raw_stderr": "1.181",
142
+ "latency": "3.478",
143
+ "cost_per_test": "0.0002",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.181,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "lcb:medium:google/gemini-1.5-flash-002:score",
163
+ "evaluation_name": "vals_ai.lcb.medium",
164
+ "source_data": {
165
+ "dataset_name": "LiveCodeBench - Medium",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/lcb"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "lcb",
172
+ "task_key": "medium",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Medium).",
179
+ "metric_id": "vals_ai.lcb.medium.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 22.977,
195
+ "details": {
196
+ "benchmark_slug": "lcb",
197
+ "benchmark_name": "LiveCodeBench",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "medium",
200
+ "task_name": "Medium",
201
+ "dataset_type": "public",
202
+ "industry": "coding",
203
+ "raw_score": "22.977",
204
+ "raw_stderr": "2.15",
205
+ "latency": "2.997",
206
+ "cost_per_test": "0.000168",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.15,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "lcb:overall:google/gemini-1.5-flash-002:score",
226
+ "evaluation_name": "vals_ai.lcb.overall",
227
+ "source_data": {
228
+ "dataset_name": "LiveCodeBench - Overall",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/lcb"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "lcb",
235
+ "task_key": "overall",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Overall).",
242
+ "metric_id": "vals_ai.lcb.overall.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 36.91,
258
+ "details": {
259
+ "benchmark_slug": "lcb",
260
+ "benchmark_name": "LiveCodeBench",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "overall",
263
+ "task_name": "Overall",
264
+ "dataset_type": "public",
265
+ "industry": "coding",
266
+ "raw_score": "36.91",
267
+ "raw_stderr": "1.079",
268
+ "latency": "2.512",
269
+ "cost_per_test": "0.000147",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.079,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ }
287
+ ]
288
+ }
data/vals_ai/google/gemini-1.5-flash-002/b4d5c708-7af5-437d-ab02-451fef628983.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/math500/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MATH 500",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "math500",
13
+ "benchmark_name": "MATH 500",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "math500:overall:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.math500.overall",
38
+ "source_data": {
39
+ "dataset_name": "MATH 500 - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/math500"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "math500",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MATH 500 (Overall).",
53
+ "metric_id": "vals_ai.math500.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 78.8,
69
+ "details": {
70
+ "benchmark_slug": "math500",
71
+ "benchmark_name": "MATH 500",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "78.8",
78
+ "raw_stderr": "3.576",
79
+ "latency": "2.645",
80
+ "cost_per_test": "0.00012",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.576,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-1.5-flash-002/bb25fb0d-5550-490b-9ebe-4cee8e8e73ea.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/corp_fin_v2/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - CorpFin v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "corp_fin_v2",
13
+ "benchmark_name": "CorpFin v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "corp_fin_v2:exact_pages:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.corp_fin_v2.exact_pages",
38
+ "source_data": {
39
+ "dataset_name": "CorpFin v2 - Exact Pages",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "corp_fin_v2",
43
+ "task_key": "exact_pages",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Exact Pages).",
50
+ "metric_id": "vals_ai.corp_fin_v2.exact_pages.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 37.762,
66
+ "details": {
67
+ "benchmark_slug": "corp_fin_v2",
68
+ "benchmark_name": "CorpFin v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "exact_pages",
71
+ "task_name": "Exact Pages",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "37.762",
75
+ "raw_stderr": "1.655",
76
+ "latency": "13.387",
77
+ "cost_per_test": "0.004915",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.655,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "corp_fin_v2:max_fitting_context:google/gemini-1.5-flash-002:score",
97
+ "evaluation_name": "vals_ai.corp_fin_v2.max_fitting_context",
98
+ "source_data": {
99
+ "dataset_name": "CorpFin v2 - Max Fitting Context",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "corp_fin_v2",
103
+ "task_key": "max_fitting_context",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Max Fitting Context).",
110
+ "metric_id": "vals_ai.corp_fin_v2.max_fitting_context.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 38.928,
126
+ "details": {
127
+ "benchmark_slug": "corp_fin_v2",
128
+ "benchmark_name": "CorpFin v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "max_fitting_context",
131
+ "task_name": "Max Fitting Context",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "38.928",
135
+ "raw_stderr": "1.665",
136
+ "latency": "58.446",
137
+ "cost_per_test": "0.013757",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.665,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "corp_fin_v2:overall:google/gemini-1.5-flash-002:score",
157
+ "evaluation_name": "vals_ai.corp_fin_v2.overall",
158
+ "source_data": {
159
+ "dataset_name": "CorpFin v2 - Overall",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "corp_fin_v2",
163
+ "task_key": "overall",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Overall).",
170
+ "metric_id": "vals_ai.corp_fin_v2.overall.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 38.19,
186
+ "details": {
187
+ "benchmark_slug": "corp_fin_v2",
188
+ "benchmark_name": "CorpFin v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "overall",
191
+ "task_name": "Overall",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "38.19",
195
+ "raw_stderr": "0.958",
196
+ "latency": "28.407",
197
+ "cost_per_test": "0.007862",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 0.958,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ },
215
+ {
216
+ "evaluation_result_id": "corp_fin_v2:shared_max_context:google/gemini-1.5-flash-002:score",
217
+ "evaluation_name": "vals_ai.corp_fin_v2.shared_max_context",
218
+ "source_data": {
219
+ "dataset_name": "CorpFin v2 - Shared Max Context",
220
+ "source_type": "other",
221
+ "additional_details": {
222
+ "benchmark_slug": "corp_fin_v2",
223
+ "task_key": "shared_max_context",
224
+ "dataset_type": "private",
225
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
226
+ }
227
+ },
228
+ "metric_config": {
229
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Shared Max Context).",
230
+ "metric_id": "vals_ai.corp_fin_v2.shared_max_context.accuracy",
231
+ "metric_name": "Accuracy",
232
+ "metric_kind": "accuracy",
233
+ "metric_unit": "percent",
234
+ "lower_is_better": false,
235
+ "score_type": "continuous",
236
+ "min_score": 0.0,
237
+ "max_score": 100.0,
238
+ "additional_details": {
239
+ "score_scale": "percent_0_to_100",
240
+ "max_score_source": "fixed_percentage_bound",
241
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
242
+ }
243
+ },
244
+ "score_details": {
245
+ "score": 37.879,
246
+ "details": {
247
+ "benchmark_slug": "corp_fin_v2",
248
+ "benchmark_name": "CorpFin v2",
249
+ "benchmark_updated": "2026-06-17",
250
+ "task_key": "shared_max_context",
251
+ "task_name": "Shared Max Context",
252
+ "dataset_type": "private",
253
+ "industry": "finance",
254
+ "raw_score": "37.879",
255
+ "raw_stderr": "1.656",
256
+ "latency": "13.387",
257
+ "cost_per_test": "0.004915",
258
+ "max_output_tokens": "8192",
259
+ "provider": "Google"
260
+ },
261
+ "uncertainty": {
262
+ "standard_error": {
263
+ "value": 1.656,
264
+ "method": "vals_reported"
265
+ }
266
+ }
267
+ },
268
+ "generation_config": {
269
+ "generation_args": {
270
+ "max_tokens": 8192,
271
+ "max_attempts": 1
272
+ }
273
+ }
274
+ }
275
+ ]
276
+ }
data/vals_ai/google/gemini-1.5-flash-002/bbf2e577-cbd3-4231-919f-6fdfc402b4a4.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmmu/google_gemini-1.5-flash-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMMU",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmmu",
13
+ "benchmark_name": "MMMU",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-flash-002",
27
+ "id": "google/gemini-1.5-flash-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-flash-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmmu:overall:google/gemini-1.5-flash-002:score",
37
+ "evaluation_name": "vals_ai.mmmu.overall",
38
+ "source_data": {
39
+ "dataset_name": "MMMU - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmmu"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmmu",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMMU (Overall).",
53
+ "metric_id": "vals_ai.mmmu.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 57.192,
69
+ "details": {
70
+ "benchmark_slug": "mmmu",
71
+ "benchmark_name": "MMMU",
72
+ "benchmark_updated": "2026-06-10",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "57.192",
78
+ "raw_stderr": "1.189",
79
+ "latency": "41.314",
80
+ "cost_per_test": "0.000114",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.189,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-1.5-pro-002/275a02e0-dccc-422d-8069-d6a521becf92.json ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/gpqa/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - GPQA Diamond",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "gpqa",
13
+ "benchmark_name": "GPQA Diamond",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "gpqa:diamond_few_shot_cot:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.gpqa.diamond_few_shot_cot",
38
+ "source_data": {
39
+ "dataset_name": "GPQA Diamond - Few-Shot CoT",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/gpqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "gpqa",
46
+ "task_key": "diamond_few_shot_cot",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Few-Shot CoT).",
53
+ "metric_id": "vals_ai.gpqa.diamond_few_shot_cot.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 56.061,
69
+ "details": {
70
+ "benchmark_slug": "gpqa",
71
+ "benchmark_name": "GPQA Diamond",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "diamond_few_shot_cot",
74
+ "task_name": "Few-Shot CoT",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "56.061",
78
+ "raw_stderr": "3.527",
79
+ "latency": "7.712",
80
+ "cost_per_test": "0.004704",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.527,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "gpqa:diamond_zero_shot_cot:google/gemini-1.5-pro-002:score",
100
+ "evaluation_name": "vals_ai.gpqa.diamond_zero_shot_cot",
101
+ "source_data": {
102
+ "dataset_name": "GPQA Diamond - Zero-Shot CoT",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/gpqa"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "gpqa",
109
+ "task_key": "diamond_zero_shot_cot",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Zero-Shot CoT).",
116
+ "metric_id": "vals_ai.gpqa.diamond_zero_shot_cot.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 60.606,
132
+ "details": {
133
+ "benchmark_slug": "gpqa",
134
+ "benchmark_name": "GPQA Diamond",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "diamond_zero_shot_cot",
137
+ "task_name": "Zero-Shot CoT",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "60.606",
141
+ "raw_stderr": "3.472",
142
+ "latency": "7.706",
143
+ "cost_per_test": "0.002606",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 3.472,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "gpqa:overall:google/gemini-1.5-pro-002:score",
163
+ "evaluation_name": "vals_ai.gpqa.overall",
164
+ "source_data": {
165
+ "dataset_name": "GPQA Diamond - Overall",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/gpqa"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "gpqa",
172
+ "task_key": "overall",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Overall).",
179
+ "metric_id": "vals_ai.gpqa.overall.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 58.334,
195
+ "details": {
196
+ "benchmark_slug": "gpqa",
197
+ "benchmark_name": "GPQA Diamond",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "overall",
200
+ "task_name": "Overall",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "58.334",
204
+ "raw_stderr": "2.475",
205
+ "latency": "7.709",
206
+ "cost_per_test": "0.003655",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.475,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ }
224
+ ]
225
+ }
data/vals_ai/google/gemini-1.5-pro-002/4f2e817f-6b6d-48b2-8ce5-4be91e29d426.json ADDED
@@ -0,0 +1,792 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mgsm/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MGSM",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mgsm",
13
+ "benchmark_name": "MGSM",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mgsm:mgsm_bn:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.mgsm.mgsm_bn",
38
+ "source_data": {
39
+ "dataset_name": "MGSM - Bengali",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mgsm"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mgsm",
46
+ "task_key": "mgsm_bn",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Bengali).",
53
+ "metric_id": "vals_ai.mgsm.mgsm_bn.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 90.4,
69
+ "details": {
70
+ "benchmark_slug": "mgsm",
71
+ "benchmark_name": "MGSM",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "mgsm_bn",
74
+ "task_name": "Bengali",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "90.4",
78
+ "raw_stderr": "1.863",
79
+ "latency": "3.331",
80
+ "cost_per_test": "0.001408",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.863,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mgsm:mgsm_de:google/gemini-1.5-pro-002:score",
100
+ "evaluation_name": "vals_ai.mgsm.mgsm_de",
101
+ "source_data": {
102
+ "dataset_name": "MGSM - German",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mgsm"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mgsm",
109
+ "task_key": "mgsm_de",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (German).",
116
+ "metric_id": "vals_ai.mgsm.mgsm_de.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 91.2,
132
+ "details": {
133
+ "benchmark_slug": "mgsm",
134
+ "benchmark_name": "MGSM",
135
+ "benchmark_updated": "2026-01-09",
136
+ "task_key": "mgsm_de",
137
+ "task_name": "German",
138
+ "dataset_type": "public",
139
+ "industry": "math",
140
+ "raw_score": "91.2",
141
+ "raw_stderr": "1.792",
142
+ "latency": "2.659",
143
+ "cost_per_test": "0.000971",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.792,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mgsm:mgsm_en:google/gemini-1.5-pro-002:score",
163
+ "evaluation_name": "vals_ai.mgsm.mgsm_en",
164
+ "source_data": {
165
+ "dataset_name": "MGSM - English",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mgsm"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mgsm",
172
+ "task_key": "mgsm_en",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (English).",
179
+ "metric_id": "vals_ai.mgsm.mgsm_en.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 94.8,
195
+ "details": {
196
+ "benchmark_slug": "mgsm",
197
+ "benchmark_name": "MGSM",
198
+ "benchmark_updated": "2026-01-09",
199
+ "task_key": "mgsm_en",
200
+ "task_name": "English",
201
+ "dataset_type": "public",
202
+ "industry": "math",
203
+ "raw_score": "94.8",
204
+ "raw_stderr": "1.404",
205
+ "latency": "1.961",
206
+ "cost_per_test": "0.000732",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.404,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mgsm:mgsm_es:google/gemini-1.5-pro-002:score",
226
+ "evaluation_name": "vals_ai.mgsm.mgsm_es",
227
+ "source_data": {
228
+ "dataset_name": "MGSM - Spanish",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mgsm"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mgsm",
235
+ "task_key": "mgsm_es",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Spanish).",
242
+ "metric_id": "vals_ai.mgsm.mgsm_es.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 92.4,
258
+ "details": {
259
+ "benchmark_slug": "mgsm",
260
+ "benchmark_name": "MGSM",
261
+ "benchmark_updated": "2026-01-09",
262
+ "task_key": "mgsm_es",
263
+ "task_name": "Spanish",
264
+ "dataset_type": "public",
265
+ "industry": "math",
266
+ "raw_score": "92.4",
267
+ "raw_stderr": "1.676",
268
+ "latency": "2.238",
269
+ "cost_per_test": "0.00082",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.676,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mgsm:mgsm_fr:google/gemini-1.5-pro-002:score",
289
+ "evaluation_name": "vals_ai.mgsm.mgsm_fr",
290
+ "source_data": {
291
+ "dataset_name": "MGSM - French",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mgsm"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mgsm",
298
+ "task_key": "mgsm_fr",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (French).",
305
+ "metric_id": "vals_ai.mgsm.mgsm_fr.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 78.8,
321
+ "details": {
322
+ "benchmark_slug": "mgsm",
323
+ "benchmark_name": "MGSM",
324
+ "benchmark_updated": "2026-01-09",
325
+ "task_key": "mgsm_fr",
326
+ "task_name": "French",
327
+ "dataset_type": "public",
328
+ "industry": "math",
329
+ "raw_score": "78.8",
330
+ "raw_stderr": "2.585",
331
+ "latency": "3.567",
332
+ "cost_per_test": "0.000934",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 2.585,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mgsm:mgsm_ja:google/gemini-1.5-pro-002:score",
352
+ "evaluation_name": "vals_ai.mgsm.mgsm_ja",
353
+ "source_data": {
354
+ "dataset_name": "MGSM - Japanese",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mgsm"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mgsm",
361
+ "task_key": "mgsm_ja",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Japanese).",
368
+ "metric_id": "vals_ai.mgsm.mgsm_ja.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 86.8,
384
+ "details": {
385
+ "benchmark_slug": "mgsm",
386
+ "benchmark_name": "MGSM",
387
+ "benchmark_updated": "2026-01-09",
388
+ "task_key": "mgsm_ja",
389
+ "task_name": "Japanese",
390
+ "dataset_type": "public",
391
+ "industry": "math",
392
+ "raw_score": "86.8",
393
+ "raw_stderr": "2.141",
394
+ "latency": "2.776",
395
+ "cost_per_test": "0.000956",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 2.141,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mgsm:mgsm_ru:google/gemini-1.5-pro-002:score",
415
+ "evaluation_name": "vals_ai.mgsm.mgsm_ru",
416
+ "source_data": {
417
+ "dataset_name": "MGSM - Russian",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mgsm"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mgsm",
424
+ "task_key": "mgsm_ru",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Russian).",
431
+ "metric_id": "vals_ai.mgsm.mgsm_ru.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 93.2,
447
+ "details": {
448
+ "benchmark_slug": "mgsm",
449
+ "benchmark_name": "MGSM",
450
+ "benchmark_updated": "2026-01-09",
451
+ "task_key": "mgsm_ru",
452
+ "task_name": "Russian",
453
+ "dataset_type": "public",
454
+ "industry": "math",
455
+ "raw_score": "93.2",
456
+ "raw_stderr": "1.592",
457
+ "latency": "2.66",
458
+ "cost_per_test": "0.000956",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.592,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mgsm:mgsm_sw:google/gemini-1.5-pro-002:score",
478
+ "evaluation_name": "vals_ai.mgsm.mgsm_sw",
479
+ "source_data": {
480
+ "dataset_name": "MGSM - Swahili",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mgsm"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mgsm",
487
+ "task_key": "mgsm_sw",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Swahili).",
494
+ "metric_id": "vals_ai.mgsm.mgsm_sw.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 88.0,
510
+ "details": {
511
+ "benchmark_slug": "mgsm",
512
+ "benchmark_name": "MGSM",
513
+ "benchmark_updated": "2026-01-09",
514
+ "task_key": "mgsm_sw",
515
+ "task_name": "Swahili",
516
+ "dataset_type": "public",
517
+ "industry": "math",
518
+ "raw_score": "88",
519
+ "raw_stderr": "2.055",
520
+ "latency": "2.513",
521
+ "cost_per_test": "0.000881",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 2.055,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mgsm:mgsm_te:google/gemini-1.5-pro-002:score",
541
+ "evaluation_name": "vals_ai.mgsm.mgsm_te",
542
+ "source_data": {
543
+ "dataset_name": "MGSM - Telugu",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mgsm"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mgsm",
550
+ "task_key": "mgsm_te",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Telugu).",
557
+ "metric_id": "vals_ai.mgsm.mgsm_te.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 86.4,
573
+ "details": {
574
+ "benchmark_slug": "mgsm",
575
+ "benchmark_name": "MGSM",
576
+ "benchmark_updated": "2026-01-09",
577
+ "task_key": "mgsm_te",
578
+ "task_name": "Telugu",
579
+ "dataset_type": "public",
580
+ "industry": "math",
581
+ "raw_score": "86.4",
582
+ "raw_stderr": "2.168",
583
+ "latency": "3.954",
584
+ "cost_per_test": "0.001663",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 2.168,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mgsm:mgsm_th:google/gemini-1.5-pro-002:score",
604
+ "evaluation_name": "vals_ai.mgsm.mgsm_th",
605
+ "source_data": {
606
+ "dataset_name": "MGSM - Thai",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mgsm"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mgsm",
613
+ "task_key": "mgsm_th",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Thai).",
620
+ "metric_id": "vals_ai.mgsm.mgsm_th.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 90.8,
636
+ "details": {
637
+ "benchmark_slug": "mgsm",
638
+ "benchmark_name": "MGSM",
639
+ "benchmark_updated": "2026-01-09",
640
+ "task_key": "mgsm_th",
641
+ "task_name": "Thai",
642
+ "dataset_type": "public",
643
+ "industry": "math",
644
+ "raw_score": "90.8",
645
+ "raw_stderr": "1.828",
646
+ "latency": "2.635",
647
+ "cost_per_test": "0.00098",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 1.828,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mgsm:mgsm_zh:google/gemini-1.5-pro-002:score",
667
+ "evaluation_name": "vals_ai.mgsm.mgsm_zh",
668
+ "source_data": {
669
+ "dataset_name": "MGSM - Chinese",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mgsm"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mgsm",
676
+ "task_key": "mgsm_zh",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Chinese).",
683
+ "metric_id": "vals_ai.mgsm.mgsm_zh.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 88.4,
699
+ "details": {
700
+ "benchmark_slug": "mgsm",
701
+ "benchmark_name": "MGSM",
702
+ "benchmark_updated": "2026-01-09",
703
+ "task_key": "mgsm_zh",
704
+ "task_name": "Chinese",
705
+ "dataset_type": "public",
706
+ "industry": "math",
707
+ "raw_score": "88.4",
708
+ "raw_stderr": "2.025",
709
+ "latency": "2.486",
710
+ "cost_per_test": "0.000854",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 2.025,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mgsm:overall:google/gemini-1.5-pro-002:score",
730
+ "evaluation_name": "vals_ai.mgsm.overall",
731
+ "source_data": {
732
+ "dataset_name": "MGSM - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mgsm"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mgsm",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Overall).",
746
+ "metric_id": "vals_ai.mgsm.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 89.2,
762
+ "details": {
763
+ "benchmark_slug": "mgsm",
764
+ "benchmark_name": "MGSM",
765
+ "benchmark_updated": "2026-01-09",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "math",
770
+ "raw_score": "89.2",
771
+ "raw_stderr": "1.921",
772
+ "latency": "2.798",
773
+ "cost_per_test": "0.001014",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 1.921,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ }
791
+ ]
792
+ }
data/vals_ai/google/gemini-1.5-pro-002/7635f61f-03d8-4879-9f76-5c870a56b2b2.json ADDED
@@ -0,0 +1,288 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/lcb/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LiveCodeBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "lcb",
13
+ "benchmark_name": "LiveCodeBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "lcb:easy:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.lcb.easy",
38
+ "source_data": {
39
+ "dataset_name": "LiveCodeBench - Easy",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/lcb"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "lcb",
46
+ "task_key": "easy",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Easy).",
53
+ "metric_id": "vals_ai.lcb.easy.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 88.199,
69
+ "details": {
70
+ "benchmark_slug": "lcb",
71
+ "benchmark_name": "LiveCodeBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "easy",
74
+ "task_name": "Easy",
75
+ "dataset_type": "public",
76
+ "industry": "coding",
77
+ "raw_score": "88.199",
78
+ "raw_stderr": "1.798",
79
+ "latency": "2.197",
80
+ "cost_per_test": "0.001189",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.798,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "lcb:hard:google/gemini-1.5-pro-002:score",
100
+ "evaluation_name": "vals_ai.lcb.hard",
101
+ "source_data": {
102
+ "dataset_name": "LiveCodeBench - Hard",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/lcb"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "lcb",
109
+ "task_key": "hard",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Hard).",
116
+ "metric_id": "vals_ai.lcb.hard.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 7.714,
132
+ "details": {
133
+ "benchmark_slug": "lcb",
134
+ "benchmark_name": "LiveCodeBench",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "hard",
137
+ "task_name": "Hard",
138
+ "dataset_type": "public",
139
+ "industry": "coding",
140
+ "raw_score": "7.714",
141
+ "raw_stderr": "1.426",
142
+ "latency": "5.667",
143
+ "cost_per_test": "0.002659",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.426,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "lcb:medium:google/gemini-1.5-pro-002:score",
163
+ "evaluation_name": "vals_ai.lcb.medium",
164
+ "source_data": {
165
+ "dataset_name": "LiveCodeBench - Medium",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/lcb"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "lcb",
172
+ "task_key": "medium",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Medium).",
179
+ "metric_id": "vals_ai.lcb.medium.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 29.243,
195
+ "details": {
196
+ "benchmark_slug": "lcb",
197
+ "benchmark_name": "LiveCodeBench",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "medium",
200
+ "task_name": "Medium",
201
+ "dataset_type": "public",
202
+ "industry": "coding",
203
+ "raw_score": "29.243",
204
+ "raw_stderr": "2.324",
205
+ "latency": "3.998",
206
+ "cost_per_test": "0.002077",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.324,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "lcb:overall:google/gemini-1.5-pro-002:score",
226
+ "evaluation_name": "vals_ai.lcb.overall",
227
+ "source_data": {
228
+ "dataset_name": "LiveCodeBench - Overall",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/lcb"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "lcb",
235
+ "task_key": "overall",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Overall).",
242
+ "metric_id": "vals_ai.lcb.overall.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 41.719,
258
+ "details": {
259
+ "benchmark_slug": "lcb",
260
+ "benchmark_name": "LiveCodeBench",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "overall",
263
+ "task_name": "Overall",
264
+ "dataset_type": "public",
265
+ "industry": "coding",
266
+ "raw_score": "41.719",
267
+ "raw_stderr": "1.089",
268
+ "latency": "3.954",
269
+ "cost_per_test": "0.001975",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.089,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ }
287
+ ]
288
+ }
data/vals_ai/google/gemini-1.5-pro-002/7e5903c6-0243-4a70-9eee-75d4c2b19c9e.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/math500/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MATH 500",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "math500",
13
+ "benchmark_name": "MATH 500",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "math500:overall:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.math500.overall",
38
+ "source_data": {
39
+ "dataset_name": "MATH 500 - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/math500"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "math500",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MATH 500 (Overall).",
53
+ "metric_id": "vals_ai.math500.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 82.8,
69
+ "details": {
70
+ "benchmark_slug": "math500",
71
+ "benchmark_name": "MATH 500",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "82.8",
78
+ "raw_stderr": "3.305",
79
+ "latency": "5.021",
80
+ "cost_per_test": "0.002164",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.305,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-1.5-pro-002/82982b53-113b-44fd-847f-ddbb3bf8dd94.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/aime/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - AIME",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "aime",
13
+ "benchmark_name": "AIME",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "aime:aime_2024:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.aime.aime_2024",
38
+ "source_data": {
39
+ "dataset_name": "AIME - AIME 2024",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/aime"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "aime",
46
+ "task_key": "aime_2024",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2024).",
53
+ "metric_id": "vals_ai.aime.aime_2024.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 23.75,
69
+ "details": {
70
+ "benchmark_slug": "aime",
71
+ "benchmark_name": "AIME",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "aime_2024",
74
+ "task_name": "AIME 2024",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "23.75",
78
+ "raw_stderr": "0.755",
79
+ "latency": "10.658",
80
+ "cost_per_test": "0.00421",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 0.755,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ }
90
+ },
91
+ {
92
+ "evaluation_result_id": "aime:aime_2025:google/gemini-1.5-pro-002:score",
93
+ "evaluation_name": "vals_ai.aime.aime_2025",
94
+ "source_data": {
95
+ "dataset_name": "AIME - AIME 2025",
96
+ "source_type": "url",
97
+ "url": [
98
+ "https://www.vals.ai/benchmarks/aime"
99
+ ],
100
+ "additional_details": {
101
+ "benchmark_slug": "aime",
102
+ "task_key": "aime_2025",
103
+ "dataset_type": "public",
104
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
105
+ }
106
+ },
107
+ "metric_config": {
108
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2025).",
109
+ "metric_id": "vals_ai.aime.aime_2025.accuracy",
110
+ "metric_name": "Accuracy",
111
+ "metric_kind": "accuracy",
112
+ "metric_unit": "percent",
113
+ "lower_is_better": false,
114
+ "score_type": "continuous",
115
+ "min_score": 0.0,
116
+ "max_score": 100.0,
117
+ "additional_details": {
118
+ "score_scale": "percent_0_to_100",
119
+ "max_score_source": "fixed_percentage_bound",
120
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
121
+ }
122
+ },
123
+ "score_details": {
124
+ "score": 13.75,
125
+ "details": {
126
+ "benchmark_slug": "aime",
127
+ "benchmark_name": "AIME",
128
+ "benchmark_updated": "2026-04-16",
129
+ "task_key": "aime_2025",
130
+ "task_name": "AIME 2025",
131
+ "dataset_type": "public",
132
+ "industry": "math",
133
+ "raw_score": "13.75",
134
+ "raw_stderr": "0.755",
135
+ "latency": "10.626",
136
+ "cost_per_test": "0.004196",
137
+ "provider": "Google"
138
+ },
139
+ "uncertainty": {
140
+ "standard_error": {
141
+ "value": 0.755,
142
+ "method": "vals_reported"
143
+ }
144
+ }
145
+ }
146
+ },
147
+ {
148
+ "evaluation_result_id": "aime:overall:google/gemini-1.5-pro-002:score",
149
+ "evaluation_name": "vals_ai.aime.overall",
150
+ "source_data": {
151
+ "dataset_name": "AIME - Overall",
152
+ "source_type": "url",
153
+ "url": [
154
+ "https://www.vals.ai/benchmarks/aime"
155
+ ],
156
+ "additional_details": {
157
+ "benchmark_slug": "aime",
158
+ "task_key": "overall",
159
+ "dataset_type": "public",
160
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
161
+ }
162
+ },
163
+ "metric_config": {
164
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (Overall).",
165
+ "metric_id": "vals_ai.aime.overall.accuracy",
166
+ "metric_name": "Accuracy",
167
+ "metric_kind": "accuracy",
168
+ "metric_unit": "percent",
169
+ "lower_is_better": false,
170
+ "score_type": "continuous",
171
+ "min_score": 0.0,
172
+ "max_score": 100.0,
173
+ "additional_details": {
174
+ "score_scale": "percent_0_to_100",
175
+ "max_score_source": "fixed_percentage_bound",
176
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
177
+ }
178
+ },
179
+ "score_details": {
180
+ "score": 18.75,
181
+ "details": {
182
+ "benchmark_slug": "aime",
183
+ "benchmark_name": "AIME",
184
+ "benchmark_updated": "2026-04-16",
185
+ "task_key": "overall",
186
+ "task_name": "Overall",
187
+ "dataset_type": "public",
188
+ "industry": "math",
189
+ "raw_score": "18.75",
190
+ "raw_stderr": "0.534",
191
+ "latency": "10.642",
192
+ "cost_per_test": "0.004203",
193
+ "provider": "Google"
194
+ },
195
+ "uncertainty": {
196
+ "standard_error": {
197
+ "value": 0.534,
198
+ "method": "vals_reported"
199
+ }
200
+ }
201
+ }
202
+ }
203
+ ]
204
+ }
data/vals_ai/google/gemini-1.5-pro-002/8a615867-e648-4856-a8e0-33c86823304f.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/corp_fin_v2/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - CorpFin v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "corp_fin_v2",
13
+ "benchmark_name": "CorpFin v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "corp_fin_v2:exact_pages:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.corp_fin_v2.exact_pages",
38
+ "source_data": {
39
+ "dataset_name": "CorpFin v2 - Exact Pages",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "corp_fin_v2",
43
+ "task_key": "exact_pages",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Exact Pages).",
50
+ "metric_id": "vals_ai.corp_fin_v2.exact_pages.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 40.676,
66
+ "details": {
67
+ "benchmark_slug": "corp_fin_v2",
68
+ "benchmark_name": "CorpFin v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "exact_pages",
71
+ "task_name": "Exact Pages",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "40.676",
75
+ "raw_stderr": "1.677",
76
+ "latency": "20.551",
77
+ "cost_per_test": "0.082015",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.677,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "corp_fin_v2:max_fitting_context:google/gemini-1.5-pro-002:score",
97
+ "evaluation_name": "vals_ai.corp_fin_v2.max_fitting_context",
98
+ "source_data": {
99
+ "dataset_name": "CorpFin v2 - Max Fitting Context",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "corp_fin_v2",
103
+ "task_key": "max_fitting_context",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Max Fitting Context).",
110
+ "metric_id": "vals_ai.corp_fin_v2.max_fitting_context.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 40.443,
126
+ "details": {
127
+ "benchmark_slug": "corp_fin_v2",
128
+ "benchmark_name": "CorpFin v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "max_fitting_context",
131
+ "task_name": "Max Fitting Context",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "40.443",
135
+ "raw_stderr": "1.675",
136
+ "latency": "73.047",
137
+ "cost_per_test": "0.229417",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.675,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "corp_fin_v2:overall:google/gemini-1.5-pro-002:score",
157
+ "evaluation_name": "vals_ai.corp_fin_v2.overall",
158
+ "source_data": {
159
+ "dataset_name": "CorpFin v2 - Overall",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "corp_fin_v2",
163
+ "task_key": "overall",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Overall).",
170
+ "metric_id": "vals_ai.corp_fin_v2.overall.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 40.521,
186
+ "details": {
187
+ "benchmark_slug": "corp_fin_v2",
188
+ "benchmark_name": "CorpFin v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "overall",
191
+ "task_name": "Overall",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "40.521",
195
+ "raw_stderr": "0.967",
196
+ "latency": "38.05",
197
+ "cost_per_test": "0.131149",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 0.967,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ },
215
+ {
216
+ "evaluation_result_id": "corp_fin_v2:shared_max_context:google/gemini-1.5-pro-002:score",
217
+ "evaluation_name": "vals_ai.corp_fin_v2.shared_max_context",
218
+ "source_data": {
219
+ "dataset_name": "CorpFin v2 - Shared Max Context",
220
+ "source_type": "other",
221
+ "additional_details": {
222
+ "benchmark_slug": "corp_fin_v2",
223
+ "task_key": "shared_max_context",
224
+ "dataset_type": "private",
225
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
226
+ }
227
+ },
228
+ "metric_config": {
229
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Shared Max Context).",
230
+ "metric_id": "vals_ai.corp_fin_v2.shared_max_context.accuracy",
231
+ "metric_name": "Accuracy",
232
+ "metric_kind": "accuracy",
233
+ "metric_unit": "percent",
234
+ "lower_is_better": false,
235
+ "score_type": "continuous",
236
+ "min_score": 0.0,
237
+ "max_score": 100.0,
238
+ "additional_details": {
239
+ "score_scale": "percent_0_to_100",
240
+ "max_score_source": "fixed_percentage_bound",
241
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
242
+ }
243
+ },
244
+ "score_details": {
245
+ "score": 40.443,
246
+ "details": {
247
+ "benchmark_slug": "corp_fin_v2",
248
+ "benchmark_name": "CorpFin v2",
249
+ "benchmark_updated": "2026-06-17",
250
+ "task_key": "shared_max_context",
251
+ "task_name": "Shared Max Context",
252
+ "dataset_type": "private",
253
+ "industry": "finance",
254
+ "raw_score": "40.443",
255
+ "raw_stderr": "1.675",
256
+ "latency": "20.551",
257
+ "cost_per_test": "0.082015",
258
+ "max_output_tokens": "8192",
259
+ "provider": "Google"
260
+ },
261
+ "uncertainty": {
262
+ "standard_error": {
263
+ "value": 1.675,
264
+ "method": "vals_reported"
265
+ }
266
+ }
267
+ },
268
+ "generation_config": {
269
+ "generation_args": {
270
+ "max_tokens": 8192,
271
+ "max_attempts": 1
272
+ }
273
+ }
274
+ }
275
+ ]
276
+ }
data/vals_ai/google/gemini-1.5-pro-002/8d504161-486f-4732-93cf-389cebc550d8.json ADDED
@@ -0,0 +1,981 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmlu_pro/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMLU Pro",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmlu_pro",
13
+ "benchmark_name": "MMLU Pro",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmlu_pro:biology:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.mmlu_pro.biology",
38
+ "source_data": {
39
+ "dataset_name": "MMLU Pro - Biology",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmlu_pro"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmlu_pro",
46
+ "task_key": "biology",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Biology).",
53
+ "metric_id": "vals_ai.mmlu_pro.biology.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 88.145,
69
+ "details": {
70
+ "benchmark_slug": "mmlu_pro",
71
+ "benchmark_name": "MMLU Pro",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "biology",
74
+ "task_name": "Biology",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "88.145",
78
+ "raw_stderr": "1.207",
79
+ "latency": "2.688",
80
+ "cost_per_test": "0.0024",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.207,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mmlu_pro:business:google/gemini-1.5-pro-002:score",
100
+ "evaluation_name": "vals_ai.mmlu_pro.business",
101
+ "source_data": {
102
+ "dataset_name": "MMLU Pro - Business",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mmlu_pro"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mmlu_pro",
109
+ "task_key": "business",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Business).",
116
+ "metric_id": "vals_ai.mmlu_pro.business.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 80.355,
132
+ "details": {
133
+ "benchmark_slug": "mmlu_pro",
134
+ "benchmark_name": "MMLU Pro",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "business",
137
+ "task_name": "Business",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "80.355",
141
+ "raw_stderr": "1.414",
142
+ "latency": "3.559",
143
+ "cost_per_test": "0.002419",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.414,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mmlu_pro:chemistry:google/gemini-1.5-pro-002:score",
163
+ "evaluation_name": "vals_ai.mmlu_pro.chemistry",
164
+ "source_data": {
165
+ "dataset_name": "MMLU Pro - Chemistry",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mmlu_pro"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mmlu_pro",
172
+ "task_key": "chemistry",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Chemistry).",
179
+ "metric_id": "vals_ai.mmlu_pro.chemistry.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 78.975,
195
+ "details": {
196
+ "benchmark_slug": "mmlu_pro",
197
+ "benchmark_name": "MMLU Pro",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "chemistry",
200
+ "task_name": "Chemistry",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "78.975",
204
+ "raw_stderr": "1.211",
205
+ "latency": "5.361",
206
+ "cost_per_test": "0.003754",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.211,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mmlu_pro:computer_science:google/gemini-1.5-pro-002:score",
226
+ "evaluation_name": "vals_ai.mmlu_pro.computer_science",
227
+ "source_data": {
228
+ "dataset_name": "MMLU Pro - Computer Science",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mmlu_pro"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mmlu_pro",
235
+ "task_key": "computer_science",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Computer Science).",
242
+ "metric_id": "vals_ai.mmlu_pro.computer_science.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 78.049,
258
+ "details": {
259
+ "benchmark_slug": "mmlu_pro",
260
+ "benchmark_name": "MMLU Pro",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "computer_science",
263
+ "task_name": "Computer Science",
264
+ "dataset_type": "public",
265
+ "industry": "academic",
266
+ "raw_score": "78.049",
267
+ "raw_stderr": "2.044",
268
+ "latency": "3.692",
269
+ "cost_per_test": "0.003045",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 2.044,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mmlu_pro:economics:google/gemini-1.5-pro-002:score",
289
+ "evaluation_name": "vals_ai.mmlu_pro.economics",
290
+ "source_data": {
291
+ "dataset_name": "MMLU Pro - Economics",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mmlu_pro"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mmlu_pro",
298
+ "task_key": "economics",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Economics).",
305
+ "metric_id": "vals_ai.mmlu_pro.economics.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 80.687,
321
+ "details": {
322
+ "benchmark_slug": "mmlu_pro",
323
+ "benchmark_name": "MMLU Pro",
324
+ "benchmark_updated": "2026-06-17",
325
+ "task_key": "economics",
326
+ "task_name": "Economics",
327
+ "dataset_type": "public",
328
+ "industry": "academic",
329
+ "raw_score": "80.687",
330
+ "raw_stderr": "1.359",
331
+ "latency": "2.351",
332
+ "cost_per_test": "0.002155",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 1.359,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mmlu_pro:engineering:google/gemini-1.5-pro-002:score",
352
+ "evaluation_name": "vals_ai.mmlu_pro.engineering",
353
+ "source_data": {
354
+ "dataset_name": "MMLU Pro - Engineering",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mmlu_pro"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mmlu_pro",
361
+ "task_key": "engineering",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Engineering).",
368
+ "metric_id": "vals_ai.mmlu_pro.engineering.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 59.752,
384
+ "details": {
385
+ "benchmark_slug": "mmlu_pro",
386
+ "benchmark_name": "MMLU Pro",
387
+ "benchmark_updated": "2026-06-17",
388
+ "task_key": "engineering",
389
+ "task_name": "Engineering",
390
+ "dataset_type": "public",
391
+ "industry": "academic",
392
+ "raw_score": "59.752",
393
+ "raw_stderr": "1.575",
394
+ "latency": "7.355",
395
+ "cost_per_test": "0.004022",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 1.575,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mmlu_pro:health:google/gemini-1.5-pro-002:score",
415
+ "evaluation_name": "vals_ai.mmlu_pro.health",
416
+ "source_data": {
417
+ "dataset_name": "MMLU Pro - Health",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mmlu_pro"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mmlu_pro",
424
+ "task_key": "health",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Health).",
431
+ "metric_id": "vals_ai.mmlu_pro.health.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 73.594,
447
+ "details": {
448
+ "benchmark_slug": "mmlu_pro",
449
+ "benchmark_name": "MMLU Pro",
450
+ "benchmark_updated": "2026-06-17",
451
+ "task_key": "health",
452
+ "task_name": "Health",
453
+ "dataset_type": "public",
454
+ "industry": "academic",
455
+ "raw_score": "73.594",
456
+ "raw_stderr": "1.544",
457
+ "latency": "2.18",
458
+ "cost_per_test": "0.001843",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.544,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mmlu_pro:history:google/gemini-1.5-pro-002:score",
478
+ "evaluation_name": "vals_ai.mmlu_pro.history",
479
+ "source_data": {
480
+ "dataset_name": "MMLU Pro - History",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mmlu_pro"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mmlu_pro",
487
+ "task_key": "history",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (History).",
494
+ "metric_id": "vals_ai.mmlu_pro.history.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 67.192,
510
+ "details": {
511
+ "benchmark_slug": "mmlu_pro",
512
+ "benchmark_name": "MMLU Pro",
513
+ "benchmark_updated": "2026-06-17",
514
+ "task_key": "history",
515
+ "task_name": "History",
516
+ "dataset_type": "public",
517
+ "industry": "academic",
518
+ "raw_score": "67.192",
519
+ "raw_stderr": "2.405",
520
+ "latency": "1.967",
521
+ "cost_per_test": "0.003125",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 2.405,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mmlu_pro:law:google/gemini-1.5-pro-002:score",
541
+ "evaluation_name": "vals_ai.mmlu_pro.law",
542
+ "source_data": {
543
+ "dataset_name": "MMLU Pro - Law",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mmlu_pro"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mmlu_pro",
550
+ "task_key": "law",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Law).",
557
+ "metric_id": "vals_ai.mmlu_pro.law.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 54.223,
573
+ "details": {
574
+ "benchmark_slug": "mmlu_pro",
575
+ "benchmark_name": "MMLU Pro",
576
+ "benchmark_updated": "2026-06-17",
577
+ "task_key": "law",
578
+ "task_name": "Law",
579
+ "dataset_type": "public",
580
+ "industry": "academic",
581
+ "raw_score": "54.223",
582
+ "raw_stderr": "1.501",
583
+ "latency": "2.483",
584
+ "cost_per_test": "0.002383",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 1.501,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mmlu_pro:math:google/gemini-1.5-pro-002:score",
604
+ "evaluation_name": "vals_ai.mmlu_pro.math",
605
+ "source_data": {
606
+ "dataset_name": "MMLU Pro - Math",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mmlu_pro"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mmlu_pro",
613
+ "task_key": "math",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Math).",
620
+ "metric_id": "vals_ai.mmlu_pro.math.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 85.936,
636
+ "details": {
637
+ "benchmark_slug": "mmlu_pro",
638
+ "benchmark_name": "MMLU Pro",
639
+ "benchmark_updated": "2026-06-17",
640
+ "task_key": "math",
641
+ "task_name": "Math",
642
+ "dataset_type": "public",
643
+ "industry": "academic",
644
+ "raw_score": "85.936",
645
+ "raw_stderr": "0.946",
646
+ "latency": "4.791",
647
+ "cost_per_test": "0.003813",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 0.946,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mmlu_pro:other:google/gemini-1.5-pro-002:score",
667
+ "evaluation_name": "vals_ai.mmlu_pro.other",
668
+ "source_data": {
669
+ "dataset_name": "MMLU Pro - Others",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mmlu_pro"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mmlu_pro",
676
+ "task_key": "other",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Others).",
683
+ "metric_id": "vals_ai.mmlu_pro.other.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 75.0,
699
+ "details": {
700
+ "benchmark_slug": "mmlu_pro",
701
+ "benchmark_name": "MMLU Pro",
702
+ "benchmark_updated": "2026-06-17",
703
+ "task_key": "other",
704
+ "task_name": "Others",
705
+ "dataset_type": "public",
706
+ "industry": "academic",
707
+ "raw_score": "75",
708
+ "raw_stderr": "1.425",
709
+ "latency": "1.857",
710
+ "cost_per_test": "0.0016",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 1.425,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mmlu_pro:overall:google/gemini-1.5-pro-002:score",
730
+ "evaluation_name": "vals_ai.mmlu_pro.overall",
731
+ "source_data": {
732
+ "dataset_name": "MMLU Pro - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mmlu_pro"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mmlu_pro",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Overall).",
746
+ "metric_id": "vals_ai.mmlu_pro.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 75.294,
762
+ "details": {
763
+ "benchmark_slug": "mmlu_pro",
764
+ "benchmark_name": "MMLU Pro",
765
+ "benchmark_updated": "2026-06-17",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "academic",
770
+ "raw_score": "75.294",
771
+ "raw_stderr": "0.417",
772
+ "latency": "3.343",
773
+ "cost_per_test": "0.002649",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 0.417,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ },
791
+ {
792
+ "evaluation_result_id": "mmlu_pro:philosophy:google/gemini-1.5-pro-002:score",
793
+ "evaluation_name": "vals_ai.mmlu_pro.philosophy",
794
+ "source_data": {
795
+ "dataset_name": "MMLU Pro - Philosophy",
796
+ "source_type": "url",
797
+ "url": [
798
+ "https://www.vals.ai/benchmarks/mmlu_pro"
799
+ ],
800
+ "additional_details": {
801
+ "benchmark_slug": "mmlu_pro",
802
+ "task_key": "philosophy",
803
+ "dataset_type": "public",
804
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
805
+ }
806
+ },
807
+ "metric_config": {
808
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Philosophy).",
809
+ "metric_id": "vals_ai.mmlu_pro.philosophy.accuracy",
810
+ "metric_name": "Accuracy",
811
+ "metric_kind": "accuracy",
812
+ "metric_unit": "percent",
813
+ "lower_is_better": false,
814
+ "score_type": "continuous",
815
+ "min_score": 0.0,
816
+ "max_score": 100.0,
817
+ "additional_details": {
818
+ "score_scale": "percent_0_to_100",
819
+ "max_score_source": "fixed_percentage_bound",
820
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
821
+ }
822
+ },
823
+ "score_details": {
824
+ "score": 70.341,
825
+ "details": {
826
+ "benchmark_slug": "mmlu_pro",
827
+ "benchmark_name": "MMLU Pro",
828
+ "benchmark_updated": "2026-06-17",
829
+ "task_key": "philosophy",
830
+ "task_name": "Philosophy",
831
+ "dataset_type": "public",
832
+ "industry": "academic",
833
+ "raw_score": "70.341",
834
+ "raw_stderr": "2.045",
835
+ "latency": "2.096",
836
+ "cost_per_test": "0.001742",
837
+ "max_output_tokens": "8192",
838
+ "provider": "Google"
839
+ },
840
+ "uncertainty": {
841
+ "standard_error": {
842
+ "value": 2.045,
843
+ "method": "vals_reported"
844
+ }
845
+ }
846
+ },
847
+ "generation_config": {
848
+ "generation_args": {
849
+ "max_tokens": 8192,
850
+ "max_attempts": 1
851
+ }
852
+ }
853
+ },
854
+ {
855
+ "evaluation_result_id": "mmlu_pro:physics:google/gemini-1.5-pro-002:score",
856
+ "evaluation_name": "vals_ai.mmlu_pro.physics",
857
+ "source_data": {
858
+ "dataset_name": "MMLU Pro - Physics",
859
+ "source_type": "url",
860
+ "url": [
861
+ "https://www.vals.ai/benchmarks/mmlu_pro"
862
+ ],
863
+ "additional_details": {
864
+ "benchmark_slug": "mmlu_pro",
865
+ "task_key": "physics",
866
+ "dataset_type": "public",
867
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
868
+ }
869
+ },
870
+ "metric_config": {
871
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Physics).",
872
+ "metric_id": "vals_ai.mmlu_pro.physics.accuracy",
873
+ "metric_name": "Accuracy",
874
+ "metric_kind": "accuracy",
875
+ "metric_unit": "percent",
876
+ "lower_is_better": false,
877
+ "score_type": "continuous",
878
+ "min_score": 0.0,
879
+ "max_score": 100.0,
880
+ "additional_details": {
881
+ "score_scale": "percent_0_to_100",
882
+ "max_score_source": "fixed_percentage_bound",
883
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
884
+ }
885
+ },
886
+ "score_details": {
887
+ "score": 81.293,
888
+ "details": {
889
+ "benchmark_slug": "mmlu_pro",
890
+ "benchmark_name": "MMLU Pro",
891
+ "benchmark_updated": "2026-06-17",
892
+ "task_key": "physics",
893
+ "task_name": "Physics",
894
+ "dataset_type": "public",
895
+ "industry": "academic",
896
+ "raw_score": "81.293",
897
+ "raw_stderr": "1.084",
898
+ "latency": "4.62",
899
+ "cost_per_test": "0.003023",
900
+ "max_output_tokens": "8192",
901
+ "provider": "Google"
902
+ },
903
+ "uncertainty": {
904
+ "standard_error": {
905
+ "value": 1.084,
906
+ "method": "vals_reported"
907
+ }
908
+ }
909
+ },
910
+ "generation_config": {
911
+ "generation_args": {
912
+ "max_tokens": 8192,
913
+ "max_attempts": 1
914
+ }
915
+ }
916
+ },
917
+ {
918
+ "evaluation_result_id": "mmlu_pro:psychology:google/gemini-1.5-pro-002:score",
919
+ "evaluation_name": "vals_ai.mmlu_pro.psychology",
920
+ "source_data": {
921
+ "dataset_name": "MMLU Pro - Psychology",
922
+ "source_type": "url",
923
+ "url": [
924
+ "https://www.vals.ai/benchmarks/mmlu_pro"
925
+ ],
926
+ "additional_details": {
927
+ "benchmark_slug": "mmlu_pro",
928
+ "task_key": "psychology",
929
+ "dataset_type": "public",
930
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
931
+ }
932
+ },
933
+ "metric_config": {
934
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Psychology).",
935
+ "metric_id": "vals_ai.mmlu_pro.psychology.accuracy",
936
+ "metric_name": "Accuracy",
937
+ "metric_kind": "accuracy",
938
+ "metric_unit": "percent",
939
+ "lower_is_better": false,
940
+ "score_type": "continuous",
941
+ "min_score": 0.0,
942
+ "max_score": 100.0,
943
+ "additional_details": {
944
+ "score_scale": "percent_0_to_100",
945
+ "max_score_source": "fixed_percentage_bound",
946
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
947
+ }
948
+ },
949
+ "score_details": {
950
+ "score": 80.576,
951
+ "details": {
952
+ "benchmark_slug": "mmlu_pro",
953
+ "benchmark_name": "MMLU Pro",
954
+ "benchmark_updated": "2026-06-17",
955
+ "task_key": "psychology",
956
+ "task_name": "Psychology",
957
+ "dataset_type": "public",
958
+ "industry": "academic",
959
+ "raw_score": "80.576",
960
+ "raw_stderr": "1.4",
961
+ "latency": "1.803",
962
+ "cost_per_test": "0.001761",
963
+ "max_output_tokens": "8192",
964
+ "provider": "Google"
965
+ },
966
+ "uncertainty": {
967
+ "standard_error": {
968
+ "value": 1.4,
969
+ "method": "vals_reported"
970
+ }
971
+ }
972
+ },
973
+ "generation_config": {
974
+ "generation_args": {
975
+ "max_tokens": 8192,
976
+ "max_attempts": 1
977
+ }
978
+ }
979
+ }
980
+ ]
981
+ }
data/vals_ai/google/gemini-1.5-pro-002/aa3c796d-e4ab-4c30-ba04-797de0a30464.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmmu/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMMU",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmmu",
13
+ "benchmark_name": "MMMU",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmmu:overall:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.mmmu.overall",
38
+ "source_data": {
39
+ "dataset_name": "MMMU - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmmu"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmmu",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMMU (Overall).",
53
+ "metric_id": "vals_ai.mmmu.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 65.511,
69
+ "details": {
70
+ "benchmark_slug": "mmmu",
71
+ "benchmark_name": "MMMU",
72
+ "benchmark_updated": "2026-06-10",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "65.511",
78
+ "raw_stderr": "1.142",
79
+ "latency": "43.736",
80
+ "cost_per_test": "0.001846",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.142,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-1.5-pro-002/abab70b0-c577-4964-93e3-7795f9c3928a.json ADDED
@@ -0,0 +1,408 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 82.32,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "82.32",
78
+ "raw_stderr": "0.88",
79
+ "latency": "0.552",
80
+ "max_output_tokens": "8192",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 0.88,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 8192,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "legal_bench:interpretation_tasks:google/gemini-1.5-pro-002:score",
99
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
100
+ "source_data": {
101
+ "dataset_name": "LegalBench - Interpretation Tasks",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/legal_bench"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "legal_bench",
108
+ "task_key": "interpretation_tasks",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
115
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 72.377,
131
+ "details": {
132
+ "benchmark_slug": "legal_bench",
133
+ "benchmark_name": "LegalBench",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "interpretation_tasks",
136
+ "task_name": "Interpretation Tasks",
137
+ "dataset_type": "public",
138
+ "industry": "legal",
139
+ "raw_score": "72.377",
140
+ "raw_stderr": "0.217",
141
+ "latency": "0.57",
142
+ "max_output_tokens": "8192",
143
+ "provider": "Google"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 0.217,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 8192,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "legal_bench:issue_tasks:google/gemini-1.5-pro-002:score",
161
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
162
+ "source_data": {
163
+ "dataset_name": "LegalBench - Issue Tasks",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/legal_bench"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "legal_bench",
170
+ "task_key": "issue_tasks",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
177
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 64.126,
193
+ "details": {
194
+ "benchmark_slug": "legal_bench",
195
+ "benchmark_name": "LegalBench",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "issue_tasks",
198
+ "task_name": "Issue Tasks",
199
+ "dataset_type": "public",
200
+ "industry": "legal",
201
+ "raw_score": "64.126",
202
+ "raw_stderr": "0.589",
203
+ "latency": "0.673",
204
+ "max_output_tokens": "8192",
205
+ "provider": "Google"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 0.589,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 8192,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "legal_bench:overall:google/gemini-1.5-pro-002:score",
223
+ "evaluation_name": "vals_ai.legal_bench.overall",
224
+ "source_data": {
225
+ "dataset_name": "LegalBench - Overall",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/legal_bench"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "legal_bench",
232
+ "task_key": "overall",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
239
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 69.08,
255
+ "details": {
256
+ "benchmark_slug": "legal_bench",
257
+ "benchmark_name": "LegalBench",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": "overall",
260
+ "task_name": "Overall",
261
+ "dataset_type": "public",
262
+ "industry": "legal",
263
+ "raw_score": "69.08",
264
+ "raw_stderr": "0.751",
265
+ "latency": "0.898",
266
+ "max_output_tokens": "8192",
267
+ "provider": "Google"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 0.751,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 8192,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ },
283
+ {
284
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:google/gemini-1.5-pro-002:score",
285
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
286
+ "source_data": {
287
+ "dataset_name": "LegalBench - Rhetoric Tasks",
288
+ "source_type": "url",
289
+ "url": [
290
+ "https://www.vals.ai/benchmarks/legal_bench"
291
+ ],
292
+ "additional_details": {
293
+ "benchmark_slug": "legal_bench",
294
+ "task_key": "rhetoric_tasks",
295
+ "dataset_type": "public",
296
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
297
+ }
298
+ },
299
+ "metric_config": {
300
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
301
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
302
+ "metric_name": "Accuracy",
303
+ "metric_kind": "accuracy",
304
+ "metric_unit": "percent",
305
+ "lower_is_better": false,
306
+ "score_type": "continuous",
307
+ "min_score": 0.0,
308
+ "max_score": 100.0,
309
+ "additional_details": {
310
+ "score_scale": "percent_0_to_100",
311
+ "max_score_source": "fixed_percentage_bound",
312
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
313
+ }
314
+ },
315
+ "score_details": {
316
+ "score": 65.108,
317
+ "details": {
318
+ "benchmark_slug": "legal_bench",
319
+ "benchmark_name": "LegalBench",
320
+ "benchmark_updated": "2026-06-17",
321
+ "task_key": "rhetoric_tasks",
322
+ "task_name": "Rhetoric Tasks",
323
+ "dataset_type": "public",
324
+ "industry": "legal",
325
+ "raw_score": "65.108",
326
+ "raw_stderr": "0.666",
327
+ "latency": "0.599",
328
+ "max_output_tokens": "8192",
329
+ "provider": "Google"
330
+ },
331
+ "uncertainty": {
332
+ "standard_error": {
333
+ "value": 0.666,
334
+ "method": "vals_reported"
335
+ }
336
+ }
337
+ },
338
+ "generation_config": {
339
+ "generation_args": {
340
+ "max_tokens": 8192,
341
+ "max_attempts": 1
342
+ }
343
+ }
344
+ },
345
+ {
346
+ "evaluation_result_id": "legal_bench:rule_tasks:google/gemini-1.5-pro-002:score",
347
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
348
+ "source_data": {
349
+ "dataset_name": "LegalBench - Rule Tasks",
350
+ "source_type": "url",
351
+ "url": [
352
+ "https://www.vals.ai/benchmarks/legal_bench"
353
+ ],
354
+ "additional_details": {
355
+ "benchmark_slug": "legal_bench",
356
+ "task_key": "rule_tasks",
357
+ "dataset_type": "public",
358
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
359
+ }
360
+ },
361
+ "metric_config": {
362
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
363
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
364
+ "metric_name": "Accuracy",
365
+ "metric_kind": "accuracy",
366
+ "metric_unit": "percent",
367
+ "lower_is_better": false,
368
+ "score_type": "continuous",
369
+ "min_score": 0.0,
370
+ "max_score": 100.0,
371
+ "additional_details": {
372
+ "score_scale": "percent_0_to_100",
373
+ "max_score_source": "fixed_percentage_bound",
374
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
375
+ }
376
+ },
377
+ "score_details": {
378
+ "score": 61.468,
379
+ "details": {
380
+ "benchmark_slug": "legal_bench",
381
+ "benchmark_name": "LegalBench",
382
+ "benchmark_updated": "2026-06-17",
383
+ "task_key": "rule_tasks",
384
+ "task_name": "Rule Tasks",
385
+ "dataset_type": "public",
386
+ "industry": "legal",
387
+ "raw_score": "61.468",
388
+ "raw_stderr": "3.533",
389
+ "latency": "2.096",
390
+ "max_output_tokens": "8192",
391
+ "provider": "Google"
392
+ },
393
+ "uncertainty": {
394
+ "standard_error": {
395
+ "value": 3.533,
396
+ "method": "vals_reported"
397
+ }
398
+ }
399
+ },
400
+ "generation_config": {
401
+ "generation_args": {
402
+ "max_tokens": 8192,
403
+ "max_attempts": 1
404
+ }
405
+ }
406
+ }
407
+ ]
408
+ }
data/vals_ai/google/gemini-1.5-pro-002/ccc83c77-cdf6-4dff-a2af-8fdfd7994da0.json ADDED
@@ -0,0 +1,477 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/medqa/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MedQA",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "medqa",
13
+ "benchmark_name": "MedQA",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "healthcare",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "medqa:asian:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.medqa.asian",
38
+ "source_data": {
39
+ "dataset_name": "MedQA - Asian",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/medqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "medqa",
46
+ "task_key": "asian",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Asian).",
53
+ "metric_id": "vals_ai.medqa.asian.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 76.45,
69
+ "details": {
70
+ "benchmark_slug": "medqa",
71
+ "benchmark_name": "MedQA",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "asian",
74
+ "task_name": "Asian",
75
+ "dataset_type": "public",
76
+ "industry": "healthcare",
77
+ "raw_score": "76.45",
78
+ "raw_stderr": "1.858",
79
+ "latency": "5.804",
80
+ "cost_per_test": "0.002938",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.858,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "medqa:black:google/gemini-1.5-pro-002:score",
100
+ "evaluation_name": "vals_ai.medqa.black",
101
+ "source_data": {
102
+ "dataset_name": "MedQA - Black",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/medqa"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "medqa",
109
+ "task_key": "black",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Black).",
116
+ "metric_id": "vals_ai.medqa.black.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 76.128,
132
+ "details": {
133
+ "benchmark_slug": "medqa",
134
+ "benchmark_name": "MedQA",
135
+ "benchmark_updated": "2026-04-16",
136
+ "task_key": "black",
137
+ "task_name": "Black",
138
+ "dataset_type": "public",
139
+ "industry": "healthcare",
140
+ "raw_score": "76.128",
141
+ "raw_stderr": "1.87",
142
+ "latency": "5.905",
143
+ "cost_per_test": "0.002995",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.87,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "medqa:hispanic:google/gemini-1.5-pro-002:score",
163
+ "evaluation_name": "vals_ai.medqa.hispanic",
164
+ "source_data": {
165
+ "dataset_name": "MedQA - Hispanic",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/medqa"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "medqa",
172
+ "task_key": "hispanic",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Hispanic).",
179
+ "metric_id": "vals_ai.medqa.hispanic.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 75.35,
195
+ "details": {
196
+ "benchmark_slug": "medqa",
197
+ "benchmark_name": "MedQA",
198
+ "benchmark_updated": "2026-04-16",
199
+ "task_key": "hispanic",
200
+ "task_name": "Hispanic",
201
+ "dataset_type": "public",
202
+ "industry": "healthcare",
203
+ "raw_score": "75.35",
204
+ "raw_stderr": "1.888",
205
+ "latency": "5.938",
206
+ "cost_per_test": "0.002973",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.888,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "medqa:indigenous:google/gemini-1.5-pro-002:score",
226
+ "evaluation_name": "vals_ai.medqa.indigenous",
227
+ "source_data": {
228
+ "dataset_name": "MedQA - Indigenous",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/medqa"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "medqa",
235
+ "task_key": "indigenous",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Indigenous).",
242
+ "metric_id": "vals_ai.medqa.indigenous.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 76.1,
258
+ "details": {
259
+ "benchmark_slug": "medqa",
260
+ "benchmark_name": "MedQA",
261
+ "benchmark_updated": "2026-04-16",
262
+ "task_key": "indigenous",
263
+ "task_name": "Indigenous",
264
+ "dataset_type": "public",
265
+ "industry": "healthcare",
266
+ "raw_score": "76.1",
267
+ "raw_stderr": "1.868",
268
+ "latency": "6.143",
269
+ "cost_per_test": "0.002998",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.868,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "medqa:overall:google/gemini-1.5-pro-002:score",
289
+ "evaluation_name": "vals_ai.medqa.overall",
290
+ "source_data": {
291
+ "dataset_name": "MedQA - Overall",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/medqa"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "medqa",
298
+ "task_key": "overall",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Overall).",
305
+ "metric_id": "vals_ai.medqa.overall.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 76.53,
321
+ "details": {
322
+ "benchmark_slug": "medqa",
323
+ "benchmark_name": "MedQA",
324
+ "benchmark_updated": "2026-04-16",
325
+ "task_key": "overall",
326
+ "task_name": "Overall",
327
+ "dataset_type": "public",
328
+ "industry": "healthcare",
329
+ "raw_score": "76.53",
330
+ "raw_stderr": "0.758",
331
+ "latency": "5.927",
332
+ "cost_per_test": "0.002954",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 0.758,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "medqa:unbiased:google/gemini-1.5-pro-002:score",
352
+ "evaluation_name": "vals_ai.medqa.unbiased",
353
+ "source_data": {
354
+ "dataset_name": "MedQA - Unbiased",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/medqa"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "medqa",
361
+ "task_key": "unbiased",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Unbiased).",
368
+ "metric_id": "vals_ai.medqa.unbiased.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 78.4,
384
+ "details": {
385
+ "benchmark_slug": "medqa",
386
+ "benchmark_name": "MedQA",
387
+ "benchmark_updated": "2026-04-16",
388
+ "task_key": "unbiased",
389
+ "task_name": "Unbiased",
390
+ "dataset_type": "public",
391
+ "industry": "healthcare",
392
+ "raw_score": "78.4",
393
+ "raw_stderr": "1.803",
394
+ "latency": "5.692",
395
+ "cost_per_test": "0.002873",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 1.803,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "medqa:white:google/gemini-1.5-pro-002:score",
415
+ "evaluation_name": "vals_ai.medqa.white",
416
+ "source_data": {
417
+ "dataset_name": "MedQA - White",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/medqa"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "medqa",
424
+ "task_key": "white",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (White).",
431
+ "metric_id": "vals_ai.medqa.white.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 76.75,
447
+ "details": {
448
+ "benchmark_slug": "medqa",
449
+ "benchmark_name": "MedQA",
450
+ "benchmark_updated": "2026-04-16",
451
+ "task_key": "white",
452
+ "task_name": "White",
453
+ "dataset_type": "public",
454
+ "industry": "healthcare",
455
+ "raw_score": "76.75",
456
+ "raw_stderr": "1.85",
457
+ "latency": "6.077",
458
+ "cost_per_test": "0.002945",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.85,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ }
476
+ ]
477
+ }
data/vals_ai/google/gemini-1.5-pro-002/dfe50d40-073b-464f-a6bd-c9ad72b68c4d.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 47.343,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "47.343",
75
+ "raw_stderr": "1.428",
76
+ "latency": "9.86",
77
+ "cost_per_test": "0.002545",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.428,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-1.5-pro-002:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 59.485,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "59.485",
135
+ "raw_stderr": "0.962",
136
+ "latency": "9.86",
137
+ "cost_per_test": "0.002545",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.962,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-1.5-pro-002:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 71.627,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "71.627",
195
+ "raw_stderr": "1.289",
196
+ "latency": "9.86",
197
+ "cost_per_test": "0.002545",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.289,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-1.5-pro-002/ff92616e-001f-403a-88d2-07ce2cf8cc95.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mortgage_tax/google_gemini-1.5-pro-002/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MortgageTax",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mortgage_tax",
13
+ "benchmark_name": "MortgageTax",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-1.5-pro-002",
27
+ "id": "google/gemini-1.5-pro-002",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-1.5-pro-002",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mortgage_tax:numerical_extraction:google/gemini-1.5-pro-002:score",
37
+ "evaluation_name": "vals_ai.mortgage_tax.numerical_extraction",
38
+ "source_data": {
39
+ "dataset_name": "MortgageTax - Numerical Extraction",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "mortgage_tax",
43
+ "task_key": "numerical_extraction",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Numerical Extraction).",
50
+ "metric_id": "vals_ai.mortgage_tax.numerical_extraction.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 63.434,
66
+ "details": {
67
+ "benchmark_slug": "mortgage_tax",
68
+ "benchmark_name": "MortgageTax",
69
+ "benchmark_updated": "2026-06-10",
70
+ "task_key": "numerical_extraction",
71
+ "task_name": "Numerical Extraction",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "63.434",
75
+ "raw_stderr": "2.658",
76
+ "latency": "3.989",
77
+ "cost_per_test": "0.001918",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 2.658,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "mortgage_tax:overall:google/gemini-1.5-pro-002:score",
97
+ "evaluation_name": "vals_ai.mortgage_tax.overall",
98
+ "source_data": {
99
+ "dataset_name": "MortgageTax - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "mortgage_tax",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Overall).",
110
+ "metric_id": "vals_ai.mortgage_tax.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 56.756,
126
+ "details": {
127
+ "benchmark_slug": "mortgage_tax",
128
+ "benchmark_name": "MortgageTax",
129
+ "benchmark_updated": "2026-06-10",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "56.756",
135
+ "raw_stderr": "1.916",
136
+ "latency": "4.205",
137
+ "cost_per_test": "0.002028",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.916,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "mortgage_tax:semantic_extraction:google/gemini-1.5-pro-002:score",
157
+ "evaluation_name": "vals_ai.mortgage_tax.semantic_extraction",
158
+ "source_data": {
159
+ "dataset_name": "MortgageTax - Semantic Extraction",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "mortgage_tax",
163
+ "task_key": "semantic_extraction",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Semantic Extraction).",
170
+ "metric_id": "vals_ai.mortgage_tax.semantic_extraction.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 50.079,
186
+ "details": {
187
+ "benchmark_slug": "mortgage_tax",
188
+ "benchmark_name": "MortgageTax",
189
+ "benchmark_updated": "2026-06-10",
190
+ "task_key": "semantic_extraction",
191
+ "task_name": "Semantic Extraction",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "50.079",
195
+ "raw_stderr": "2.759",
196
+ "latency": "4.421",
197
+ "cost_per_test": "0.002137",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 2.759,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-001/14c3feef-f7fc-49a4-b370-e10c78646af7.json ADDED
@@ -0,0 +1,408 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/legal_bench/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LegalBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "legal_bench",
13
+ "benchmark_name": "LegalBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "legal",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "legal_bench:conclusion_tasks:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.legal_bench.conclusion_tasks",
38
+ "source_data": {
39
+ "dataset_name": "LegalBench - Conclusion Tasks",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/legal_bench"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "legal_bench",
46
+ "task_key": "conclusion_tasks",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Conclusion Tasks).",
53
+ "metric_id": "vals_ai.legal_bench.conclusion_tasks.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 80.388,
69
+ "details": {
70
+ "benchmark_slug": "legal_bench",
71
+ "benchmark_name": "LegalBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "conclusion_tasks",
74
+ "task_name": "Conclusion Tasks",
75
+ "dataset_type": "public",
76
+ "industry": "legal",
77
+ "raw_score": "80.388",
78
+ "raw_stderr": "0.956",
79
+ "latency": "0.316",
80
+ "max_output_tokens": "8192",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 0.956,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ },
90
+ "generation_config": {
91
+ "generation_args": {
92
+ "max_tokens": 8192,
93
+ "max_attempts": 1
94
+ }
95
+ }
96
+ },
97
+ {
98
+ "evaluation_result_id": "legal_bench:interpretation_tasks:google/gemini-2.0-flash-001:score",
99
+ "evaluation_name": "vals_ai.legal_bench.interpretation_tasks",
100
+ "source_data": {
101
+ "dataset_name": "LegalBench - Interpretation Tasks",
102
+ "source_type": "url",
103
+ "url": [
104
+ "https://www.vals.ai/benchmarks/legal_bench"
105
+ ],
106
+ "additional_details": {
107
+ "benchmark_slug": "legal_bench",
108
+ "task_key": "interpretation_tasks",
109
+ "dataset_type": "public",
110
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
111
+ }
112
+ },
113
+ "metric_config": {
114
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Interpretation Tasks).",
115
+ "metric_id": "vals_ai.legal_bench.interpretation_tasks.accuracy",
116
+ "metric_name": "Accuracy",
117
+ "metric_kind": "accuracy",
118
+ "metric_unit": "percent",
119
+ "lower_is_better": false,
120
+ "score_type": "continuous",
121
+ "min_score": 0.0,
122
+ "max_score": 100.0,
123
+ "additional_details": {
124
+ "score_scale": "percent_0_to_100",
125
+ "max_score_source": "fixed_percentage_bound",
126
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
127
+ }
128
+ },
129
+ "score_details": {
130
+ "score": 78.938,
131
+ "details": {
132
+ "benchmark_slug": "legal_bench",
133
+ "benchmark_name": "LegalBench",
134
+ "benchmark_updated": "2026-06-17",
135
+ "task_key": "interpretation_tasks",
136
+ "task_name": "Interpretation Tasks",
137
+ "dataset_type": "public",
138
+ "industry": "legal",
139
+ "raw_score": "78.938",
140
+ "raw_stderr": "0.203",
141
+ "latency": "0.322",
142
+ "max_output_tokens": "8192",
143
+ "provider": "Google"
144
+ },
145
+ "uncertainty": {
146
+ "standard_error": {
147
+ "value": 0.203,
148
+ "method": "vals_reported"
149
+ }
150
+ }
151
+ },
152
+ "generation_config": {
153
+ "generation_args": {
154
+ "max_tokens": 8192,
155
+ "max_attempts": 1
156
+ }
157
+ }
158
+ },
159
+ {
160
+ "evaluation_result_id": "legal_bench:issue_tasks:google/gemini-2.0-flash-001:score",
161
+ "evaluation_name": "vals_ai.legal_bench.issue_tasks",
162
+ "source_data": {
163
+ "dataset_name": "LegalBench - Issue Tasks",
164
+ "source_type": "url",
165
+ "url": [
166
+ "https://www.vals.ai/benchmarks/legal_bench"
167
+ ],
168
+ "additional_details": {
169
+ "benchmark_slug": "legal_bench",
170
+ "task_key": "issue_tasks",
171
+ "dataset_type": "public",
172
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
173
+ }
174
+ },
175
+ "metric_config": {
176
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Issue Tasks).",
177
+ "metric_id": "vals_ai.legal_bench.issue_tasks.accuracy",
178
+ "metric_name": "Accuracy",
179
+ "metric_kind": "accuracy",
180
+ "metric_unit": "percent",
181
+ "lower_is_better": false,
182
+ "score_type": "continuous",
183
+ "min_score": 0.0,
184
+ "max_score": 100.0,
185
+ "additional_details": {
186
+ "score_scale": "percent_0_to_100",
187
+ "max_score_source": "fixed_percentage_bound",
188
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
189
+ }
190
+ },
191
+ "score_details": {
192
+ "score": 87.905,
193
+ "details": {
194
+ "benchmark_slug": "legal_bench",
195
+ "benchmark_name": "LegalBench",
196
+ "benchmark_updated": "2026-06-17",
197
+ "task_key": "issue_tasks",
198
+ "task_name": "Issue Tasks",
199
+ "dataset_type": "public",
200
+ "industry": "legal",
201
+ "raw_score": "87.905",
202
+ "raw_stderr": "0.585",
203
+ "latency": "0.334",
204
+ "max_output_tokens": "8192",
205
+ "provider": "Google"
206
+ },
207
+ "uncertainty": {
208
+ "standard_error": {
209
+ "value": 0.585,
210
+ "method": "vals_reported"
211
+ }
212
+ }
213
+ },
214
+ "generation_config": {
215
+ "generation_args": {
216
+ "max_tokens": 8192,
217
+ "max_attempts": 1
218
+ }
219
+ }
220
+ },
221
+ {
222
+ "evaluation_result_id": "legal_bench:overall:google/gemini-2.0-flash-001:score",
223
+ "evaluation_name": "vals_ai.legal_bench.overall",
224
+ "source_data": {
225
+ "dataset_name": "LegalBench - Overall",
226
+ "source_type": "url",
227
+ "url": [
228
+ "https://www.vals.ai/benchmarks/legal_bench"
229
+ ],
230
+ "additional_details": {
231
+ "benchmark_slug": "legal_bench",
232
+ "task_key": "overall",
233
+ "dataset_type": "public",
234
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
235
+ }
236
+ },
237
+ "metric_config": {
238
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Overall).",
239
+ "metric_id": "vals_ai.legal_bench.overall.accuracy",
240
+ "metric_name": "Accuracy",
241
+ "metric_kind": "accuracy",
242
+ "metric_unit": "percent",
243
+ "lower_is_better": false,
244
+ "score_type": "continuous",
245
+ "min_score": 0.0,
246
+ "max_score": 100.0,
247
+ "additional_details": {
248
+ "score_scale": "percent_0_to_100",
249
+ "max_score_source": "fixed_percentage_bound",
250
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
251
+ }
252
+ },
253
+ "score_details": {
254
+ "score": 78.364,
255
+ "details": {
256
+ "benchmark_slug": "legal_bench",
257
+ "benchmark_name": "LegalBench",
258
+ "benchmark_updated": "2026-06-17",
259
+ "task_key": "overall",
260
+ "task_name": "Overall",
261
+ "dataset_type": "public",
262
+ "industry": "legal",
263
+ "raw_score": "78.364",
264
+ "raw_stderr": "0.782",
265
+ "latency": "0.447",
266
+ "max_output_tokens": "8192",
267
+ "provider": "Google"
268
+ },
269
+ "uncertainty": {
270
+ "standard_error": {
271
+ "value": 0.782,
272
+ "method": "vals_reported"
273
+ }
274
+ }
275
+ },
276
+ "generation_config": {
277
+ "generation_args": {
278
+ "max_tokens": 8192,
279
+ "max_attempts": 1
280
+ }
281
+ }
282
+ },
283
+ {
284
+ "evaluation_result_id": "legal_bench:rhetoric_tasks:google/gemini-2.0-flash-001:score",
285
+ "evaluation_name": "vals_ai.legal_bench.rhetoric_tasks",
286
+ "source_data": {
287
+ "dataset_name": "LegalBench - Rhetoric Tasks",
288
+ "source_type": "url",
289
+ "url": [
290
+ "https://www.vals.ai/benchmarks/legal_bench"
291
+ ],
292
+ "additional_details": {
293
+ "benchmark_slug": "legal_bench",
294
+ "task_key": "rhetoric_tasks",
295
+ "dataset_type": "public",
296
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
297
+ }
298
+ },
299
+ "metric_config": {
300
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rhetoric Tasks).",
301
+ "metric_id": "vals_ai.legal_bench.rhetoric_tasks.accuracy",
302
+ "metric_name": "Accuracy",
303
+ "metric_kind": "accuracy",
304
+ "metric_unit": "percent",
305
+ "lower_is_better": false,
306
+ "score_type": "continuous",
307
+ "min_score": 0.0,
308
+ "max_score": 100.0,
309
+ "additional_details": {
310
+ "score_scale": "percent_0_to_100",
311
+ "max_score_source": "fixed_percentage_bound",
312
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
313
+ }
314
+ },
315
+ "score_details": {
316
+ "score": 80.211,
317
+ "details": {
318
+ "benchmark_slug": "legal_bench",
319
+ "benchmark_name": "LegalBench",
320
+ "benchmark_updated": "2026-06-17",
321
+ "task_key": "rhetoric_tasks",
322
+ "task_name": "Rhetoric Tasks",
323
+ "dataset_type": "public",
324
+ "industry": "legal",
325
+ "raw_score": "80.211",
326
+ "raw_stderr": "1.04",
327
+ "latency": "0.324",
328
+ "max_output_tokens": "8192",
329
+ "provider": "Google"
330
+ },
331
+ "uncertainty": {
332
+ "standard_error": {
333
+ "value": 1.04,
334
+ "method": "vals_reported"
335
+ }
336
+ }
337
+ },
338
+ "generation_config": {
339
+ "generation_args": {
340
+ "max_tokens": 8192,
341
+ "max_attempts": 1
342
+ }
343
+ }
344
+ },
345
+ {
346
+ "evaluation_result_id": "legal_bench:rule_tasks:google/gemini-2.0-flash-001:score",
347
+ "evaluation_name": "vals_ai.legal_bench.rule_tasks",
348
+ "source_data": {
349
+ "dataset_name": "LegalBench - Rule Tasks",
350
+ "source_type": "url",
351
+ "url": [
352
+ "https://www.vals.ai/benchmarks/legal_bench"
353
+ ],
354
+ "additional_details": {
355
+ "benchmark_slug": "legal_bench",
356
+ "task_key": "rule_tasks",
357
+ "dataset_type": "public",
358
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
359
+ }
360
+ },
361
+ "metric_config": {
362
+ "evaluation_description": "Accuracy reported by Vals.ai for LegalBench (Rule Tasks).",
363
+ "metric_id": "vals_ai.legal_bench.rule_tasks.accuracy",
364
+ "metric_name": "Accuracy",
365
+ "metric_kind": "accuracy",
366
+ "metric_unit": "percent",
367
+ "lower_is_better": false,
368
+ "score_type": "continuous",
369
+ "min_score": 0.0,
370
+ "max_score": 100.0,
371
+ "additional_details": {
372
+ "score_scale": "percent_0_to_100",
373
+ "max_score_source": "fixed_percentage_bound",
374
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/legal_bench"
375
+ }
376
+ },
377
+ "score_details": {
378
+ "score": 64.379,
379
+ "details": {
380
+ "benchmark_slug": "legal_bench",
381
+ "benchmark_name": "LegalBench",
382
+ "benchmark_updated": "2026-06-17",
383
+ "task_key": "rule_tasks",
384
+ "task_name": "Rule Tasks",
385
+ "dataset_type": "public",
386
+ "industry": "legal",
387
+ "raw_score": "64.379",
388
+ "raw_stderr": "3.591",
389
+ "latency": "0.939",
390
+ "max_output_tokens": "8192",
391
+ "provider": "Google"
392
+ },
393
+ "uncertainty": {
394
+ "standard_error": {
395
+ "value": 3.591,
396
+ "method": "vals_reported"
397
+ }
398
+ }
399
+ },
400
+ "generation_config": {
401
+ "generation_args": {
402
+ "max_tokens": 8192,
403
+ "max_attempts": 1
404
+ }
405
+ }
406
+ }
407
+ ]
408
+ }
data/vals_ai/google/gemini-2.0-flash-001/32c87288-dbd0-4b0c-b10f-615110169253.json ADDED
@@ -0,0 +1,288 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/lcb/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - LiveCodeBench",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "lcb",
13
+ "benchmark_name": "LiveCodeBench",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "coding",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "lcb:easy:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.lcb.easy",
38
+ "source_data": {
39
+ "dataset_name": "LiveCodeBench - Easy",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/lcb"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "lcb",
46
+ "task_key": "easy",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Easy).",
53
+ "metric_id": "vals_ai.lcb.easy.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 90.683,
69
+ "details": {
70
+ "benchmark_slug": "lcb",
71
+ "benchmark_name": "LiveCodeBench",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "easy",
74
+ "task_name": "Easy",
75
+ "dataset_type": "public",
76
+ "industry": "coding",
77
+ "raw_score": "90.683",
78
+ "raw_stderr": "1.62",
79
+ "latency": "1.313",
80
+ "cost_per_test": "0.000123",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.62,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "lcb:hard:google/gemini-2.0-flash-001:score",
100
+ "evaluation_name": "vals_ai.lcb.hard",
101
+ "source_data": {
102
+ "dataset_name": "LiveCodeBench - Hard",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/lcb"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "lcb",
109
+ "task_key": "hard",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Hard).",
116
+ "metric_id": "vals_ai.lcb.hard.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 8.286,
132
+ "details": {
133
+ "benchmark_slug": "lcb",
134
+ "benchmark_name": "LiveCodeBench",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "hard",
137
+ "task_name": "Hard",
138
+ "dataset_type": "public",
139
+ "industry": "coding",
140
+ "raw_score": "8.286",
141
+ "raw_stderr": "1.473",
142
+ "latency": "5.372",
143
+ "cost_per_test": "0.000482",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.473,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "lcb:medium:google/gemini-2.0-flash-001:score",
163
+ "evaluation_name": "vals_ai.lcb.medium",
164
+ "source_data": {
165
+ "dataset_name": "LiveCodeBench - Medium",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/lcb"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "lcb",
172
+ "task_key": "medium",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Medium).",
179
+ "metric_id": "vals_ai.lcb.medium.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 31.854,
195
+ "details": {
196
+ "benchmark_slug": "lcb",
197
+ "benchmark_name": "LiveCodeBench",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "medium",
200
+ "task_name": "Medium",
201
+ "dataset_type": "public",
202
+ "industry": "coding",
203
+ "raw_score": "31.854",
204
+ "raw_stderr": "2.381",
205
+ "latency": "3.388",
206
+ "cost_per_test": "0.000308",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.381,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "lcb:overall:google/gemini-2.0-flash-001:score",
226
+ "evaluation_name": "vals_ai.lcb.overall",
227
+ "source_data": {
228
+ "dataset_name": "LiveCodeBench - Overall",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/lcb"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "lcb",
235
+ "task_key": "overall",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for LiveCodeBench (Overall).",
242
+ "metric_id": "vals_ai.lcb.overall.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/lcb"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 43.608,
258
+ "details": {
259
+ "benchmark_slug": "lcb",
260
+ "benchmark_name": "LiveCodeBench",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "overall",
263
+ "task_name": "Overall",
264
+ "dataset_type": "public",
265
+ "industry": "coding",
266
+ "raw_score": "43.608",
267
+ "raw_stderr": "1.078",
268
+ "latency": "3.358",
269
+ "cost_per_test": "0.000304",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.078,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ }
287
+ ]
288
+ }
data/vals_ai/google/gemini-2.0-flash-001/39454daa-ad55-452f-a9fd-30f95d80ddf8.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 53.475,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "53.475",
75
+ "raw_stderr": "1.426",
76
+ "latency": "5.702",
77
+ "cost_per_test": "0.000296",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.426,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-2.0-flash-001:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 65.25,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "65.25",
135
+ "raw_stderr": "0.933",
136
+ "latency": "5.702",
137
+ "cost_per_test": "0.000296",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.933,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-2.0-flash-001:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 77.024,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "77.024",
195
+ "raw_stderr": "1.203",
196
+ "latency": "5.702",
197
+ "cost_per_test": "0.000296",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.203,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-001/3f54213b-2b24-4a9c-a3d6-1c2c16a52060.json ADDED
@@ -0,0 +1,225 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/gpqa/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - GPQA Diamond",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "gpqa",
13
+ "benchmark_name": "GPQA Diamond",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "gpqa:diamond_few_shot_cot:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.gpqa.diamond_few_shot_cot",
38
+ "source_data": {
39
+ "dataset_name": "GPQA Diamond - Few-Shot CoT",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/gpqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "gpqa",
46
+ "task_key": "diamond_few_shot_cot",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Few-Shot CoT).",
53
+ "metric_id": "vals_ai.gpqa.diamond_few_shot_cot.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 66.162,
69
+ "details": {
70
+ "benchmark_slug": "gpqa",
71
+ "benchmark_name": "GPQA Diamond",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "diamond_few_shot_cot",
74
+ "task_name": "Few-Shot CoT",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "66.162",
78
+ "raw_stderr": "3.363",
79
+ "latency": "4.976",
80
+ "cost_per_test": "0.000512",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.363,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "gpqa:diamond_zero_shot_cot:google/gemini-2.0-flash-001:score",
100
+ "evaluation_name": "vals_ai.gpqa.diamond_zero_shot_cot",
101
+ "source_data": {
102
+ "dataset_name": "GPQA Diamond - Zero-Shot CoT",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/gpqa"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "gpqa",
109
+ "task_key": "diamond_zero_shot_cot",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Zero-Shot CoT).",
116
+ "metric_id": "vals_ai.gpqa.diamond_zero_shot_cot.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 64.141,
132
+ "details": {
133
+ "benchmark_slug": "gpqa",
134
+ "benchmark_name": "GPQA Diamond",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "diamond_zero_shot_cot",
137
+ "task_name": "Zero-Shot CoT",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "64.141",
141
+ "raw_stderr": "3.408",
142
+ "latency": "6.63",
143
+ "cost_per_test": "0.000469",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 3.408,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "gpqa:overall:google/gemini-2.0-flash-001:score",
163
+ "evaluation_name": "vals_ai.gpqa.overall",
164
+ "source_data": {
165
+ "dataset_name": "GPQA Diamond - Overall",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/gpqa"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "gpqa",
172
+ "task_key": "overall",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for GPQA Diamond (Overall).",
179
+ "metric_id": "vals_ai.gpqa.overall.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/gpqa"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 65.152,
195
+ "details": {
196
+ "benchmark_slug": "gpqa",
197
+ "benchmark_name": "GPQA Diamond",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "overall",
200
+ "task_name": "Overall",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "65.152",
204
+ "raw_stderr": "2.394",
205
+ "latency": "5.803",
206
+ "cost_per_test": "0.00049",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 2.394,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ }
224
+ ]
225
+ }
data/vals_ai/google/gemini-2.0-flash-001/6780858a-b0a5-4d52-b273-a01126dd746c.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmmu/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMMU",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmmu",
13
+ "benchmark_name": "MMMU",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmmu:overall:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.mmmu.overall",
38
+ "source_data": {
39
+ "dataset_name": "MMMU - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmmu"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmmu",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMMU (Overall).",
53
+ "metric_id": "vals_ai.mmmu.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmmu"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 69.786,
69
+ "details": {
70
+ "benchmark_slug": "mmmu",
71
+ "benchmark_name": "MMMU",
72
+ "benchmark_updated": "2026-06-10",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "69.786",
78
+ "raw_stderr": "1.104",
79
+ "latency": "45.203",
80
+ "cost_per_test": "0.000365",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.104,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-2.0-flash-001/6bd9c901-bd97-4f47-9dda-811857eb95ab.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mortgage_tax/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MortgageTax",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mortgage_tax",
13
+ "benchmark_name": "MortgageTax",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mortgage_tax:numerical_extraction:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.mortgage_tax.numerical_extraction",
38
+ "source_data": {
39
+ "dataset_name": "MortgageTax - Numerical Extraction",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "mortgage_tax",
43
+ "task_key": "numerical_extraction",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Numerical Extraction).",
50
+ "metric_id": "vals_ai.mortgage_tax.numerical_extraction.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 62.957,
66
+ "details": {
67
+ "benchmark_slug": "mortgage_tax",
68
+ "benchmark_name": "MortgageTax",
69
+ "benchmark_updated": "2026-06-10",
70
+ "task_key": "numerical_extraction",
71
+ "task_name": "Numerical Extraction",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "62.957",
75
+ "raw_stderr": "2.665",
76
+ "latency": "2.918",
77
+ "cost_per_test": "0.000265",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 2.665,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "mortgage_tax:overall:google/gemini-2.0-flash-001:score",
97
+ "evaluation_name": "vals_ai.mortgage_tax.overall",
98
+ "source_data": {
99
+ "dataset_name": "MortgageTax - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "mortgage_tax",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Overall).",
110
+ "metric_id": "vals_ai.mortgage_tax.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 59.658,
126
+ "details": {
127
+ "benchmark_slug": "mortgage_tax",
128
+ "benchmark_name": "MortgageTax",
129
+ "benchmark_updated": "2026-06-10",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "59.658",
135
+ "raw_stderr": "1.91",
136
+ "latency": "3.794",
137
+ "cost_per_test": "0.00027",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.91,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "mortgage_tax:semantic_extraction:google/gemini-2.0-flash-001:score",
157
+ "evaluation_name": "vals_ai.mortgage_tax.semantic_extraction",
158
+ "source_data": {
159
+ "dataset_name": "MortgageTax - Semantic Extraction",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "mortgage_tax",
163
+ "task_key": "semantic_extraction",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Semantic Extraction).",
170
+ "metric_id": "vals_ai.mortgage_tax.semantic_extraction.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 56.359,
186
+ "details": {
187
+ "benchmark_slug": "mortgage_tax",
188
+ "benchmark_name": "MortgageTax",
189
+ "benchmark_updated": "2026-06-10",
190
+ "task_key": "semantic_extraction",
191
+ "task_name": "Semantic Extraction",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "56.359",
195
+ "raw_stderr": "2.736",
196
+ "latency": "4.671",
197
+ "cost_per_test": "0.000274",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 2.736,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-001/85735689-b835-41fe-ba96-c2987c64e3da.json ADDED
@@ -0,0 +1,276 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/corp_fin_v2/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - CorpFin v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "corp_fin_v2",
13
+ "benchmark_name": "CorpFin v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "corp_fin_v2:exact_pages:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.corp_fin_v2.exact_pages",
38
+ "source_data": {
39
+ "dataset_name": "CorpFin v2 - Exact Pages",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "corp_fin_v2",
43
+ "task_key": "exact_pages",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Exact Pages).",
50
+ "metric_id": "vals_ai.corp_fin_v2.exact_pages.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 35.431,
66
+ "details": {
67
+ "benchmark_slug": "corp_fin_v2",
68
+ "benchmark_name": "CorpFin v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "exact_pages",
71
+ "task_name": "Exact Pages",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "35.431",
75
+ "raw_stderr": "1.633",
76
+ "latency": "18.434",
77
+ "cost_per_test": "0.006552",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.633,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "corp_fin_v2:max_fitting_context:google/gemini-2.0-flash-001:score",
97
+ "evaluation_name": "vals_ai.corp_fin_v2.max_fitting_context",
98
+ "source_data": {
99
+ "dataset_name": "CorpFin v2 - Max Fitting Context",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "corp_fin_v2",
103
+ "task_key": "max_fitting_context",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Max Fitting Context).",
110
+ "metric_id": "vals_ai.corp_fin_v2.max_fitting_context.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 30.07,
126
+ "details": {
127
+ "benchmark_slug": "corp_fin_v2",
128
+ "benchmark_name": "CorpFin v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "max_fitting_context",
131
+ "task_name": "Max Fitting Context",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "30.07",
135
+ "raw_stderr": "1.566",
136
+ "latency": "53.247",
137
+ "cost_per_test": "0.018339",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.566,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "corp_fin_v2:overall:google/gemini-2.0-flash-001:score",
157
+ "evaluation_name": "vals_ai.corp_fin_v2.overall",
158
+ "source_data": {
159
+ "dataset_name": "CorpFin v2 - Overall",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "corp_fin_v2",
163
+ "task_key": "overall",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Overall).",
170
+ "metric_id": "vals_ai.corp_fin_v2.overall.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 33.722,
186
+ "details": {
187
+ "benchmark_slug": "corp_fin_v2",
188
+ "benchmark_name": "CorpFin v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "overall",
191
+ "task_name": "Overall",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "33.722",
195
+ "raw_stderr": "0.93",
196
+ "latency": "30.038",
197
+ "cost_per_test": "0.010481",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 0.93,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ },
215
+ {
216
+ "evaluation_result_id": "corp_fin_v2:shared_max_context:google/gemini-2.0-flash-001:score",
217
+ "evaluation_name": "vals_ai.corp_fin_v2.shared_max_context",
218
+ "source_data": {
219
+ "dataset_name": "CorpFin v2 - Shared Max Context",
220
+ "source_type": "other",
221
+ "additional_details": {
222
+ "benchmark_slug": "corp_fin_v2",
223
+ "task_key": "shared_max_context",
224
+ "dataset_type": "private",
225
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
226
+ }
227
+ },
228
+ "metric_config": {
229
+ "evaluation_description": "Accuracy reported by Vals.ai for CorpFin v2 (Shared Max Context).",
230
+ "metric_id": "vals_ai.corp_fin_v2.shared_max_context.accuracy",
231
+ "metric_name": "Accuracy",
232
+ "metric_kind": "accuracy",
233
+ "metric_unit": "percent",
234
+ "lower_is_better": false,
235
+ "score_type": "continuous",
236
+ "min_score": 0.0,
237
+ "max_score": 100.0,
238
+ "additional_details": {
239
+ "score_scale": "percent_0_to_100",
240
+ "max_score_source": "fixed_percentage_bound",
241
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/corp_fin_v2"
242
+ }
243
+ },
244
+ "score_details": {
245
+ "score": 35.664,
246
+ "details": {
247
+ "benchmark_slug": "corp_fin_v2",
248
+ "benchmark_name": "CorpFin v2",
249
+ "benchmark_updated": "2026-06-17",
250
+ "task_key": "shared_max_context",
251
+ "task_name": "Shared Max Context",
252
+ "dataset_type": "private",
253
+ "industry": "finance",
254
+ "raw_score": "35.664",
255
+ "raw_stderr": "1.635",
256
+ "latency": "18.434",
257
+ "cost_per_test": "0.006552",
258
+ "max_output_tokens": "8192",
259
+ "provider": "Google"
260
+ },
261
+ "uncertainty": {
262
+ "standard_error": {
263
+ "value": 1.635,
264
+ "method": "vals_reported"
265
+ }
266
+ }
267
+ },
268
+ "generation_config": {
269
+ "generation_args": {
270
+ "max_tokens": 8192,
271
+ "max_attempts": 1
272
+ }
273
+ }
274
+ }
275
+ ]
276
+ }
data/vals_ai/google/gemini-2.0-flash-001/9e750eb8-5989-42df-ab2e-89febc4e4e7c.json ADDED
@@ -0,0 +1,792 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mgsm/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MGSM",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mgsm",
13
+ "benchmark_name": "MGSM",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mgsm:mgsm_bn:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.mgsm.mgsm_bn",
38
+ "source_data": {
39
+ "dataset_name": "MGSM - Bengali",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mgsm"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mgsm",
46
+ "task_key": "mgsm_bn",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Bengali).",
53
+ "metric_id": "vals_ai.mgsm.mgsm_bn.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 70.8,
69
+ "details": {
70
+ "benchmark_slug": "mgsm",
71
+ "benchmark_name": "MGSM",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "mgsm_bn",
74
+ "task_name": "Bengali",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "70.8",
78
+ "raw_stderr": "2.876",
79
+ "latency": "1.512",
80
+ "cost_per_test": "0.000133",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 2.876,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mgsm:mgsm_de:google/gemini-2.0-flash-001:score",
100
+ "evaluation_name": "vals_ai.mgsm.mgsm_de",
101
+ "source_data": {
102
+ "dataset_name": "MGSM - German",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mgsm"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mgsm",
109
+ "task_key": "mgsm_de",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (German).",
116
+ "metric_id": "vals_ai.mgsm.mgsm_de.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 91.6,
132
+ "details": {
133
+ "benchmark_slug": "mgsm",
134
+ "benchmark_name": "MGSM",
135
+ "benchmark_updated": "2026-01-09",
136
+ "task_key": "mgsm_de",
137
+ "task_name": "German",
138
+ "dataset_type": "public",
139
+ "industry": "math",
140
+ "raw_score": "91.6",
141
+ "raw_stderr": "1.754",
142
+ "latency": "1.497",
143
+ "cost_per_test": "8.8e-05",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.754,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mgsm:mgsm_en:google/gemini-2.0-flash-001:score",
163
+ "evaluation_name": "vals_ai.mgsm.mgsm_en",
164
+ "source_data": {
165
+ "dataset_name": "MGSM - English",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mgsm"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mgsm",
172
+ "task_key": "mgsm_en",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (English).",
179
+ "metric_id": "vals_ai.mgsm.mgsm_en.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 95.6,
195
+ "details": {
196
+ "benchmark_slug": "mgsm",
197
+ "benchmark_name": "MGSM",
198
+ "benchmark_updated": "2026-01-09",
199
+ "task_key": "mgsm_en",
200
+ "task_name": "English",
201
+ "dataset_type": "public",
202
+ "industry": "math",
203
+ "raw_score": "95.6",
204
+ "raw_stderr": "1.297",
205
+ "latency": "1.38",
206
+ "cost_per_test": "8.9e-05",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.297,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mgsm:mgsm_es:google/gemini-2.0-flash-001:score",
226
+ "evaluation_name": "vals_ai.mgsm.mgsm_es",
227
+ "source_data": {
228
+ "dataset_name": "MGSM - Spanish",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mgsm"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mgsm",
235
+ "task_key": "mgsm_es",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Spanish).",
242
+ "metric_id": "vals_ai.mgsm.mgsm_es.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 92.8,
258
+ "details": {
259
+ "benchmark_slug": "mgsm",
260
+ "benchmark_name": "MGSM",
261
+ "benchmark_updated": "2026-01-09",
262
+ "task_key": "mgsm_es",
263
+ "task_name": "Spanish",
264
+ "dataset_type": "public",
265
+ "industry": "math",
266
+ "raw_score": "92.8",
267
+ "raw_stderr": "1.635",
268
+ "latency": "1.707",
269
+ "cost_per_test": "0.000109",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.635,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mgsm:mgsm_fr:google/gemini-2.0-flash-001:score",
289
+ "evaluation_name": "vals_ai.mgsm.mgsm_fr",
290
+ "source_data": {
291
+ "dataset_name": "MGSM - French",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mgsm"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mgsm",
298
+ "task_key": "mgsm_fr",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (French).",
305
+ "metric_id": "vals_ai.mgsm.mgsm_fr.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 83.2,
321
+ "details": {
322
+ "benchmark_slug": "mgsm",
323
+ "benchmark_name": "MGSM",
324
+ "benchmark_updated": "2026-01-09",
325
+ "task_key": "mgsm_fr",
326
+ "task_name": "French",
327
+ "dataset_type": "public",
328
+ "industry": "math",
329
+ "raw_score": "83.2",
330
+ "raw_stderr": "2.364",
331
+ "latency": "1.874",
332
+ "cost_per_test": "0.00012",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 2.364,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mgsm:mgsm_ja:google/gemini-2.0-flash-001:score",
352
+ "evaluation_name": "vals_ai.mgsm.mgsm_ja",
353
+ "source_data": {
354
+ "dataset_name": "MGSM - Japanese",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mgsm"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mgsm",
361
+ "task_key": "mgsm_ja",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Japanese).",
368
+ "metric_id": "vals_ai.mgsm.mgsm_ja.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 87.2,
384
+ "details": {
385
+ "benchmark_slug": "mgsm",
386
+ "benchmark_name": "MGSM",
387
+ "benchmark_updated": "2026-01-09",
388
+ "task_key": "mgsm_ja",
389
+ "task_name": "Japanese",
390
+ "dataset_type": "public",
391
+ "industry": "math",
392
+ "raw_score": "87.2",
393
+ "raw_stderr": "2.113",
394
+ "latency": "1.366",
395
+ "cost_per_test": "7.6e-05",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 2.113,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mgsm:mgsm_ru:google/gemini-2.0-flash-001:score",
415
+ "evaluation_name": "vals_ai.mgsm.mgsm_ru",
416
+ "source_data": {
417
+ "dataset_name": "MGSM - Russian",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mgsm"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mgsm",
424
+ "task_key": "mgsm_ru",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Russian).",
431
+ "metric_id": "vals_ai.mgsm.mgsm_ru.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 96.0,
447
+ "details": {
448
+ "benchmark_slug": "mgsm",
449
+ "benchmark_name": "MGSM",
450
+ "benchmark_updated": "2026-01-09",
451
+ "task_key": "mgsm_ru",
452
+ "task_name": "Russian",
453
+ "dataset_type": "public",
454
+ "industry": "math",
455
+ "raw_score": "96",
456
+ "raw_stderr": "1.239",
457
+ "latency": "1.558",
458
+ "cost_per_test": "9.3e-05",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.239,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mgsm:mgsm_sw:google/gemini-2.0-flash-001:score",
478
+ "evaluation_name": "vals_ai.mgsm.mgsm_sw",
479
+ "source_data": {
480
+ "dataset_name": "MGSM - Swahili",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mgsm"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mgsm",
487
+ "task_key": "mgsm_sw",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Swahili).",
494
+ "metric_id": "vals_ai.mgsm.mgsm_sw.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 90.4,
510
+ "details": {
511
+ "benchmark_slug": "mgsm",
512
+ "benchmark_name": "MGSM",
513
+ "benchmark_updated": "2026-01-09",
514
+ "task_key": "mgsm_sw",
515
+ "task_name": "Swahili",
516
+ "dataset_type": "public",
517
+ "industry": "math",
518
+ "raw_score": "90.4",
519
+ "raw_stderr": "1.863",
520
+ "latency": "1.483",
521
+ "cost_per_test": "9.8e-05",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 1.863,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mgsm:mgsm_te:google/gemini-2.0-flash-001:score",
541
+ "evaluation_name": "vals_ai.mgsm.mgsm_te",
542
+ "source_data": {
543
+ "dataset_name": "MGSM - Telugu",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mgsm"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mgsm",
550
+ "task_key": "mgsm_te",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Telugu).",
557
+ "metric_id": "vals_ai.mgsm.mgsm_te.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 87.6,
573
+ "details": {
574
+ "benchmark_slug": "mgsm",
575
+ "benchmark_name": "MGSM",
576
+ "benchmark_updated": "2026-01-09",
577
+ "task_key": "mgsm_te",
578
+ "task_name": "Telugu",
579
+ "dataset_type": "public",
580
+ "industry": "math",
581
+ "raw_score": "87.6",
582
+ "raw_stderr": "2.084",
583
+ "latency": "1.788",
584
+ "cost_per_test": "0.000142",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 2.084,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mgsm:mgsm_th:google/gemini-2.0-flash-001:score",
604
+ "evaluation_name": "vals_ai.mgsm.mgsm_th",
605
+ "source_data": {
606
+ "dataset_name": "MGSM - Thai",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mgsm"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mgsm",
613
+ "task_key": "mgsm_th",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Thai).",
620
+ "metric_id": "vals_ai.mgsm.mgsm_th.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 93.2,
636
+ "details": {
637
+ "benchmark_slug": "mgsm",
638
+ "benchmark_name": "MGSM",
639
+ "benchmark_updated": "2026-01-09",
640
+ "task_key": "mgsm_th",
641
+ "task_name": "Thai",
642
+ "dataset_type": "public",
643
+ "industry": "math",
644
+ "raw_score": "93.2",
645
+ "raw_stderr": "1.592",
646
+ "latency": "1.352",
647
+ "cost_per_test": "9.1e-05",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 1.592,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mgsm:mgsm_zh:google/gemini-2.0-flash-001:score",
667
+ "evaluation_name": "vals_ai.mgsm.mgsm_zh",
668
+ "source_data": {
669
+ "dataset_name": "MGSM - Chinese",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mgsm"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mgsm",
676
+ "task_key": "mgsm_zh",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Chinese).",
683
+ "metric_id": "vals_ai.mgsm.mgsm_zh.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 90.8,
699
+ "details": {
700
+ "benchmark_slug": "mgsm",
701
+ "benchmark_name": "MGSM",
702
+ "benchmark_updated": "2026-01-09",
703
+ "task_key": "mgsm_zh",
704
+ "task_name": "Chinese",
705
+ "dataset_type": "public",
706
+ "industry": "math",
707
+ "raw_score": "90.8",
708
+ "raw_stderr": "1.828",
709
+ "latency": "1.382",
710
+ "cost_per_test": "8e-05",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 1.828,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mgsm:overall:google/gemini-2.0-flash-001:score",
730
+ "evaluation_name": "vals_ai.mgsm.overall",
731
+ "source_data": {
732
+ "dataset_name": "MGSM - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mgsm"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mgsm",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MGSM (Overall).",
746
+ "metric_id": "vals_ai.mgsm.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mgsm"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 89.018,
762
+ "details": {
763
+ "benchmark_slug": "mgsm",
764
+ "benchmark_name": "MGSM",
765
+ "benchmark_updated": "2026-01-09",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "math",
770
+ "raw_score": "89.018",
771
+ "raw_stderr": "1.877",
772
+ "latency": "1.536",
773
+ "cost_per_test": "0.000102",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 1.877,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ }
791
+ ]
792
+ }
data/vals_ai/google/gemini-2.0-flash-001/a98d1b8f-32ce-4bb4-a010-55f3e1af3612.json ADDED
@@ -0,0 +1,981 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mmlu_pro/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MMLU Pro",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mmlu_pro",
13
+ "benchmark_name": "MMLU Pro",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "public",
16
+ "industry": "academic",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mmlu_pro:biology:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.mmlu_pro.biology",
38
+ "source_data": {
39
+ "dataset_name": "MMLU Pro - Biology",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/mmlu_pro"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "mmlu_pro",
46
+ "task_key": "biology",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Biology).",
53
+ "metric_id": "vals_ai.mmlu_pro.biology.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 88.703,
69
+ "details": {
70
+ "benchmark_slug": "mmlu_pro",
71
+ "benchmark_name": "MMLU Pro",
72
+ "benchmark_updated": "2026-06-17",
73
+ "task_key": "biology",
74
+ "task_name": "Biology",
75
+ "dataset_type": "public",
76
+ "industry": "academic",
77
+ "raw_score": "88.703",
78
+ "raw_stderr": "1.182",
79
+ "latency": "2.946",
80
+ "cost_per_test": "0.000309",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.182,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "mmlu_pro:business:google/gemini-2.0-flash-001:score",
100
+ "evaluation_name": "vals_ai.mmlu_pro.business",
101
+ "source_data": {
102
+ "dataset_name": "MMLU Pro - Business",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/mmlu_pro"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "mmlu_pro",
109
+ "task_key": "business",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Business).",
116
+ "metric_id": "vals_ai.mmlu_pro.business.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 81.876,
132
+ "details": {
133
+ "benchmark_slug": "mmlu_pro",
134
+ "benchmark_name": "MMLU Pro",
135
+ "benchmark_updated": "2026-06-17",
136
+ "task_key": "business",
137
+ "task_name": "Business",
138
+ "dataset_type": "public",
139
+ "industry": "academic",
140
+ "raw_score": "81.876",
141
+ "raw_stderr": "1.371",
142
+ "latency": "4.674",
143
+ "cost_per_test": "0.000434",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.371,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "mmlu_pro:chemistry:google/gemini-2.0-flash-001:score",
163
+ "evaluation_name": "vals_ai.mmlu_pro.chemistry",
164
+ "source_data": {
165
+ "dataset_name": "MMLU Pro - Chemistry",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/mmlu_pro"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "mmlu_pro",
172
+ "task_key": "chemistry",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Chemistry).",
179
+ "metric_id": "vals_ai.mmlu_pro.chemistry.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 83.127,
195
+ "details": {
196
+ "benchmark_slug": "mmlu_pro",
197
+ "benchmark_name": "MMLU Pro",
198
+ "benchmark_updated": "2026-06-17",
199
+ "task_key": "chemistry",
200
+ "task_name": "Chemistry",
201
+ "dataset_type": "public",
202
+ "industry": "academic",
203
+ "raw_score": "83.127",
204
+ "raw_stderr": "1.113",
205
+ "latency": "6.964",
206
+ "cost_per_test": "0.000652",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.113,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "mmlu_pro:computer_science:google/gemini-2.0-flash-001:score",
226
+ "evaluation_name": "vals_ai.mmlu_pro.computer_science",
227
+ "source_data": {
228
+ "dataset_name": "MMLU Pro - Computer Science",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/mmlu_pro"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "mmlu_pro",
235
+ "task_key": "computer_science",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Computer Science).",
242
+ "metric_id": "vals_ai.mmlu_pro.computer_science.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 80.488,
258
+ "details": {
259
+ "benchmark_slug": "mmlu_pro",
260
+ "benchmark_name": "MMLU Pro",
261
+ "benchmark_updated": "2026-06-17",
262
+ "task_key": "computer_science",
263
+ "task_name": "Computer Science",
264
+ "dataset_type": "public",
265
+ "industry": "academic",
266
+ "raw_score": "80.488",
267
+ "raw_stderr": "1.957",
268
+ "latency": "6.226",
269
+ "cost_per_test": "0.000593",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.957,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "mmlu_pro:economics:google/gemini-2.0-flash-001:score",
289
+ "evaluation_name": "vals_ai.mmlu_pro.economics",
290
+ "source_data": {
291
+ "dataset_name": "MMLU Pro - Economics",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/mmlu_pro"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "mmlu_pro",
298
+ "task_key": "economics",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Economics).",
305
+ "metric_id": "vals_ai.mmlu_pro.economics.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 83.649,
321
+ "details": {
322
+ "benchmark_slug": "mmlu_pro",
323
+ "benchmark_name": "MMLU Pro",
324
+ "benchmark_updated": "2026-06-17",
325
+ "task_key": "economics",
326
+ "task_name": "Economics",
327
+ "dataset_type": "public",
328
+ "industry": "academic",
329
+ "raw_score": "83.649",
330
+ "raw_stderr": "1.273",
331
+ "latency": "2.017",
332
+ "cost_per_test": "0.000231",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 1.273,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "mmlu_pro:engineering:google/gemini-2.0-flash-001:score",
352
+ "evaluation_name": "vals_ai.mmlu_pro.engineering",
353
+ "source_data": {
354
+ "dataset_name": "MMLU Pro - Engineering",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/mmlu_pro"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "mmlu_pro",
361
+ "task_key": "engineering",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Engineering).",
368
+ "metric_id": "vals_ai.mmlu_pro.engineering.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 67.699,
384
+ "details": {
385
+ "benchmark_slug": "mmlu_pro",
386
+ "benchmark_name": "MMLU Pro",
387
+ "benchmark_updated": "2026-06-17",
388
+ "task_key": "engineering",
389
+ "task_name": "Engineering",
390
+ "dataset_type": "public",
391
+ "industry": "academic",
392
+ "raw_score": "67.699",
393
+ "raw_stderr": "1.502",
394
+ "latency": "9.302",
395
+ "cost_per_test": "0.000771",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 1.502,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "mmlu_pro:health:google/gemini-2.0-flash-001:score",
415
+ "evaluation_name": "vals_ai.mmlu_pro.health",
416
+ "source_data": {
417
+ "dataset_name": "MMLU Pro - Health",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/mmlu_pro"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "mmlu_pro",
424
+ "task_key": "health",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Health).",
431
+ "metric_id": "vals_ai.mmlu_pro.health.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 75.917,
447
+ "details": {
448
+ "benchmark_slug": "mmlu_pro",
449
+ "benchmark_name": "MMLU Pro",
450
+ "benchmark_updated": "2026-06-17",
451
+ "task_key": "health",
452
+ "task_name": "Health",
453
+ "dataset_type": "public",
454
+ "industry": "academic",
455
+ "raw_score": "75.917",
456
+ "raw_stderr": "1.495",
457
+ "latency": "2.117",
458
+ "cost_per_test": "0.000209",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.495,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ },
476
+ {
477
+ "evaluation_result_id": "mmlu_pro:history:google/gemini-2.0-flash-001:score",
478
+ "evaluation_name": "vals_ai.mmlu_pro.history",
479
+ "source_data": {
480
+ "dataset_name": "MMLU Pro - History",
481
+ "source_type": "url",
482
+ "url": [
483
+ "https://www.vals.ai/benchmarks/mmlu_pro"
484
+ ],
485
+ "additional_details": {
486
+ "benchmark_slug": "mmlu_pro",
487
+ "task_key": "history",
488
+ "dataset_type": "public",
489
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
490
+ }
491
+ },
492
+ "metric_config": {
493
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (History).",
494
+ "metric_id": "vals_ai.mmlu_pro.history.accuracy",
495
+ "metric_name": "Accuracy",
496
+ "metric_kind": "accuracy",
497
+ "metric_unit": "percent",
498
+ "lower_is_better": false,
499
+ "score_type": "continuous",
500
+ "min_score": 0.0,
501
+ "max_score": 100.0,
502
+ "additional_details": {
503
+ "score_scale": "percent_0_to_100",
504
+ "max_score_source": "fixed_percentage_bound",
505
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
506
+ }
507
+ },
508
+ "score_details": {
509
+ "score": 69.029,
510
+ "details": {
511
+ "benchmark_slug": "mmlu_pro",
512
+ "benchmark_name": "MMLU Pro",
513
+ "benchmark_updated": "2026-06-17",
514
+ "task_key": "history",
515
+ "task_name": "History",
516
+ "dataset_type": "public",
517
+ "industry": "academic",
518
+ "raw_score": "69.029",
519
+ "raw_stderr": "2.369",
520
+ "latency": "1.28",
521
+ "cost_per_test": "0.000268",
522
+ "max_output_tokens": "8192",
523
+ "provider": "Google"
524
+ },
525
+ "uncertainty": {
526
+ "standard_error": {
527
+ "value": 2.369,
528
+ "method": "vals_reported"
529
+ }
530
+ }
531
+ },
532
+ "generation_config": {
533
+ "generation_args": {
534
+ "max_tokens": 8192,
535
+ "max_attempts": 1
536
+ }
537
+ }
538
+ },
539
+ {
540
+ "evaluation_result_id": "mmlu_pro:law:google/gemini-2.0-flash-001:score",
541
+ "evaluation_name": "vals_ai.mmlu_pro.law",
542
+ "source_data": {
543
+ "dataset_name": "MMLU Pro - Law",
544
+ "source_type": "url",
545
+ "url": [
546
+ "https://www.vals.ai/benchmarks/mmlu_pro"
547
+ ],
548
+ "additional_details": {
549
+ "benchmark_slug": "mmlu_pro",
550
+ "task_key": "law",
551
+ "dataset_type": "public",
552
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
553
+ }
554
+ },
555
+ "metric_config": {
556
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Law).",
557
+ "metric_id": "vals_ai.mmlu_pro.law.accuracy",
558
+ "metric_name": "Accuracy",
559
+ "metric_kind": "accuracy",
560
+ "metric_unit": "percent",
561
+ "lower_is_better": false,
562
+ "score_type": "continuous",
563
+ "min_score": 0.0,
564
+ "max_score": 100.0,
565
+ "additional_details": {
566
+ "score_scale": "percent_0_to_100",
567
+ "max_score_source": "fixed_percentage_bound",
568
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
569
+ }
570
+ },
571
+ "score_details": {
572
+ "score": 60.672,
573
+ "details": {
574
+ "benchmark_slug": "mmlu_pro",
575
+ "benchmark_name": "MMLU Pro",
576
+ "benchmark_updated": "2026-06-17",
577
+ "task_key": "law",
578
+ "task_name": "Law",
579
+ "dataset_type": "public",
580
+ "industry": "academic",
581
+ "raw_score": "60.672",
582
+ "raw_stderr": "1.472",
583
+ "latency": "5.07",
584
+ "cost_per_test": "0.000399",
585
+ "max_output_tokens": "8192",
586
+ "provider": "Google"
587
+ },
588
+ "uncertainty": {
589
+ "standard_error": {
590
+ "value": 1.472,
591
+ "method": "vals_reported"
592
+ }
593
+ }
594
+ },
595
+ "generation_config": {
596
+ "generation_args": {
597
+ "max_tokens": 8192,
598
+ "max_attempts": 1
599
+ }
600
+ }
601
+ },
602
+ {
603
+ "evaluation_result_id": "mmlu_pro:math:google/gemini-2.0-flash-001:score",
604
+ "evaluation_name": "vals_ai.mmlu_pro.math",
605
+ "source_data": {
606
+ "dataset_name": "MMLU Pro - Math",
607
+ "source_type": "url",
608
+ "url": [
609
+ "https://www.vals.ai/benchmarks/mmlu_pro"
610
+ ],
611
+ "additional_details": {
612
+ "benchmark_slug": "mmlu_pro",
613
+ "task_key": "math",
614
+ "dataset_type": "public",
615
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
616
+ }
617
+ },
618
+ "metric_config": {
619
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Math).",
620
+ "metric_id": "vals_ai.mmlu_pro.math.accuracy",
621
+ "metric_name": "Accuracy",
622
+ "metric_kind": "accuracy",
623
+ "metric_unit": "percent",
624
+ "lower_is_better": false,
625
+ "score_type": "continuous",
626
+ "min_score": 0.0,
627
+ "max_score": 100.0,
628
+ "additional_details": {
629
+ "score_scale": "percent_0_to_100",
630
+ "max_score_source": "fixed_percentage_bound",
631
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
632
+ }
633
+ },
634
+ "score_details": {
635
+ "score": 80.755,
636
+ "details": {
637
+ "benchmark_slug": "mmlu_pro",
638
+ "benchmark_name": "MMLU Pro",
639
+ "benchmark_updated": "2026-06-17",
640
+ "task_key": "math",
641
+ "task_name": "Math",
642
+ "dataset_type": "public",
643
+ "industry": "academic",
644
+ "raw_score": "80.755",
645
+ "raw_stderr": "1.073",
646
+ "latency": "8.408",
647
+ "cost_per_test": "0.000825",
648
+ "max_output_tokens": "8192",
649
+ "provider": "Google"
650
+ },
651
+ "uncertainty": {
652
+ "standard_error": {
653
+ "value": 1.073,
654
+ "method": "vals_reported"
655
+ }
656
+ }
657
+ },
658
+ "generation_config": {
659
+ "generation_args": {
660
+ "max_tokens": 8192,
661
+ "max_attempts": 1
662
+ }
663
+ }
664
+ },
665
+ {
666
+ "evaluation_result_id": "mmlu_pro:other:google/gemini-2.0-flash-001:score",
667
+ "evaluation_name": "vals_ai.mmlu_pro.other",
668
+ "source_data": {
669
+ "dataset_name": "MMLU Pro - Others",
670
+ "source_type": "url",
671
+ "url": [
672
+ "https://www.vals.ai/benchmarks/mmlu_pro"
673
+ ],
674
+ "additional_details": {
675
+ "benchmark_slug": "mmlu_pro",
676
+ "task_key": "other",
677
+ "dataset_type": "public",
678
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
679
+ }
680
+ },
681
+ "metric_config": {
682
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Others).",
683
+ "metric_id": "vals_ai.mmlu_pro.other.accuracy",
684
+ "metric_name": "Accuracy",
685
+ "metric_kind": "accuracy",
686
+ "metric_unit": "percent",
687
+ "lower_is_better": false,
688
+ "score_type": "continuous",
689
+ "min_score": 0.0,
690
+ "max_score": 100.0,
691
+ "additional_details": {
692
+ "score_scale": "percent_0_to_100",
693
+ "max_score_source": "fixed_percentage_bound",
694
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
695
+ }
696
+ },
697
+ "score_details": {
698
+ "score": 75.325,
699
+ "details": {
700
+ "benchmark_slug": "mmlu_pro",
701
+ "benchmark_name": "MMLU Pro",
702
+ "benchmark_updated": "2026-06-17",
703
+ "task_key": "other",
704
+ "task_name": "Others",
705
+ "dataset_type": "public",
706
+ "industry": "academic",
707
+ "raw_score": "75.325",
708
+ "raw_stderr": "1.418",
709
+ "latency": "1.702",
710
+ "cost_per_test": "0.000172",
711
+ "max_output_tokens": "8192",
712
+ "provider": "Google"
713
+ },
714
+ "uncertainty": {
715
+ "standard_error": {
716
+ "value": 1.418,
717
+ "method": "vals_reported"
718
+ }
719
+ }
720
+ },
721
+ "generation_config": {
722
+ "generation_args": {
723
+ "max_tokens": 8192,
724
+ "max_attempts": 1
725
+ }
726
+ }
727
+ },
728
+ {
729
+ "evaluation_result_id": "mmlu_pro:overall:google/gemini-2.0-flash-001:score",
730
+ "evaluation_name": "vals_ai.mmlu_pro.overall",
731
+ "source_data": {
732
+ "dataset_name": "MMLU Pro - Overall",
733
+ "source_type": "url",
734
+ "url": [
735
+ "https://www.vals.ai/benchmarks/mmlu_pro"
736
+ ],
737
+ "additional_details": {
738
+ "benchmark_slug": "mmlu_pro",
739
+ "task_key": "overall",
740
+ "dataset_type": "public",
741
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
742
+ }
743
+ },
744
+ "metric_config": {
745
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Overall).",
746
+ "metric_id": "vals_ai.mmlu_pro.overall.accuracy",
747
+ "metric_name": "Accuracy",
748
+ "metric_kind": "accuracy",
749
+ "metric_unit": "percent",
750
+ "lower_is_better": false,
751
+ "score_type": "continuous",
752
+ "min_score": 0.0,
753
+ "max_score": 100.0,
754
+ "additional_details": {
755
+ "score_scale": "percent_0_to_100",
756
+ "max_score_source": "fixed_percentage_bound",
757
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
758
+ }
759
+ },
760
+ "score_details": {
761
+ "score": 77.376,
762
+ "details": {
763
+ "benchmark_slug": "mmlu_pro",
764
+ "benchmark_name": "MMLU Pro",
765
+ "benchmark_updated": "2026-06-17",
766
+ "task_key": "overall",
767
+ "task_name": "Overall",
768
+ "dataset_type": "public",
769
+ "industry": "academic",
770
+ "raw_score": "77.376",
771
+ "raw_stderr": "0.407",
772
+ "latency": "4.318",
773
+ "cost_per_test": "0.000414",
774
+ "max_output_tokens": "8192",
775
+ "provider": "Google"
776
+ },
777
+ "uncertainty": {
778
+ "standard_error": {
779
+ "value": 0.407,
780
+ "method": "vals_reported"
781
+ }
782
+ }
783
+ },
784
+ "generation_config": {
785
+ "generation_args": {
786
+ "max_tokens": 8192,
787
+ "max_attempts": 1
788
+ }
789
+ }
790
+ },
791
+ {
792
+ "evaluation_result_id": "mmlu_pro:philosophy:google/gemini-2.0-flash-001:score",
793
+ "evaluation_name": "vals_ai.mmlu_pro.philosophy",
794
+ "source_data": {
795
+ "dataset_name": "MMLU Pro - Philosophy",
796
+ "source_type": "url",
797
+ "url": [
798
+ "https://www.vals.ai/benchmarks/mmlu_pro"
799
+ ],
800
+ "additional_details": {
801
+ "benchmark_slug": "mmlu_pro",
802
+ "task_key": "philosophy",
803
+ "dataset_type": "public",
804
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
805
+ }
806
+ },
807
+ "metric_config": {
808
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Philosophy).",
809
+ "metric_id": "vals_ai.mmlu_pro.philosophy.accuracy",
810
+ "metric_name": "Accuracy",
811
+ "metric_kind": "accuracy",
812
+ "metric_unit": "percent",
813
+ "lower_is_better": false,
814
+ "score_type": "continuous",
815
+ "min_score": 0.0,
816
+ "max_score": 100.0,
817
+ "additional_details": {
818
+ "score_scale": "percent_0_to_100",
819
+ "max_score_source": "fixed_percentage_bound",
820
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
821
+ }
822
+ },
823
+ "score_details": {
824
+ "score": 71.543,
825
+ "details": {
826
+ "benchmark_slug": "mmlu_pro",
827
+ "benchmark_name": "MMLU Pro",
828
+ "benchmark_updated": "2026-06-17",
829
+ "task_key": "philosophy",
830
+ "task_name": "Philosophy",
831
+ "dataset_type": "public",
832
+ "industry": "academic",
833
+ "raw_score": "71.543",
834
+ "raw_stderr": "2.02",
835
+ "latency": "2.256",
836
+ "cost_per_test": "0.00022",
837
+ "max_output_tokens": "8192",
838
+ "provider": "Google"
839
+ },
840
+ "uncertainty": {
841
+ "standard_error": {
842
+ "value": 2.02,
843
+ "method": "vals_reported"
844
+ }
845
+ }
846
+ },
847
+ "generation_config": {
848
+ "generation_args": {
849
+ "max_tokens": 8192,
850
+ "max_attempts": 1
851
+ }
852
+ }
853
+ },
854
+ {
855
+ "evaluation_result_id": "mmlu_pro:physics:google/gemini-2.0-flash-001:score",
856
+ "evaluation_name": "vals_ai.mmlu_pro.physics",
857
+ "source_data": {
858
+ "dataset_name": "MMLU Pro - Physics",
859
+ "source_type": "url",
860
+ "url": [
861
+ "https://www.vals.ai/benchmarks/mmlu_pro"
862
+ ],
863
+ "additional_details": {
864
+ "benchmark_slug": "mmlu_pro",
865
+ "task_key": "physics",
866
+ "dataset_type": "public",
867
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
868
+ }
869
+ },
870
+ "metric_config": {
871
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Physics).",
872
+ "metric_id": "vals_ai.mmlu_pro.physics.accuracy",
873
+ "metric_name": "Accuracy",
874
+ "metric_kind": "accuracy",
875
+ "metric_unit": "percent",
876
+ "lower_is_better": false,
877
+ "score_type": "continuous",
878
+ "min_score": 0.0,
879
+ "max_score": 100.0,
880
+ "additional_details": {
881
+ "score_scale": "percent_0_to_100",
882
+ "max_score_source": "fixed_percentage_bound",
883
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
884
+ }
885
+ },
886
+ "score_details": {
887
+ "score": 83.911,
888
+ "details": {
889
+ "benchmark_slug": "mmlu_pro",
890
+ "benchmark_name": "MMLU Pro",
891
+ "benchmark_updated": "2026-06-17",
892
+ "task_key": "physics",
893
+ "task_name": "Physics",
894
+ "dataset_type": "public",
895
+ "industry": "academic",
896
+ "raw_score": "83.911",
897
+ "raw_stderr": "1.019",
898
+ "latency": "5.669",
899
+ "cost_per_test": "0.000517",
900
+ "max_output_tokens": "8192",
901
+ "provider": "Google"
902
+ },
903
+ "uncertainty": {
904
+ "standard_error": {
905
+ "value": 1.019,
906
+ "method": "vals_reported"
907
+ }
908
+ }
909
+ },
910
+ "generation_config": {
911
+ "generation_args": {
912
+ "max_tokens": 8192,
913
+ "max_attempts": 1
914
+ }
915
+ }
916
+ },
917
+ {
918
+ "evaluation_result_id": "mmlu_pro:psychology:google/gemini-2.0-flash-001:score",
919
+ "evaluation_name": "vals_ai.mmlu_pro.psychology",
920
+ "source_data": {
921
+ "dataset_name": "MMLU Pro - Psychology",
922
+ "source_type": "url",
923
+ "url": [
924
+ "https://www.vals.ai/benchmarks/mmlu_pro"
925
+ ],
926
+ "additional_details": {
927
+ "benchmark_slug": "mmlu_pro",
928
+ "task_key": "psychology",
929
+ "dataset_type": "public",
930
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
931
+ }
932
+ },
933
+ "metric_config": {
934
+ "evaluation_description": "Accuracy reported by Vals.ai for MMLU Pro (Psychology).",
935
+ "metric_id": "vals_ai.mmlu_pro.psychology.accuracy",
936
+ "metric_name": "Accuracy",
937
+ "metric_kind": "accuracy",
938
+ "metric_unit": "percent",
939
+ "lower_is_better": false,
940
+ "score_type": "continuous",
941
+ "min_score": 0.0,
942
+ "max_score": 100.0,
943
+ "additional_details": {
944
+ "score_scale": "percent_0_to_100",
945
+ "max_score_source": "fixed_percentage_bound",
946
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mmlu_pro"
947
+ }
948
+ },
949
+ "score_details": {
950
+ "score": 80.576,
951
+ "details": {
952
+ "benchmark_slug": "mmlu_pro",
953
+ "benchmark_name": "MMLU Pro",
954
+ "benchmark_updated": "2026-06-17",
955
+ "task_key": "psychology",
956
+ "task_name": "Psychology",
957
+ "dataset_type": "public",
958
+ "industry": "academic",
959
+ "raw_score": "80.576",
960
+ "raw_stderr": "1.404",
961
+ "latency": "1.826",
962
+ "cost_per_test": "0.000194",
963
+ "max_output_tokens": "8192",
964
+ "provider": "Google"
965
+ },
966
+ "uncertainty": {
967
+ "standard_error": {
968
+ "value": 1.404,
969
+ "method": "vals_reported"
970
+ }
971
+ }
972
+ },
973
+ "generation_config": {
974
+ "generation_args": {
975
+ "max_tokens": 8192,
976
+ "max_attempts": 1
977
+ }
978
+ }
979
+ }
980
+ ]
981
+ }
data/vals_ai/google/gemini-2.0-flash-001/c14ecf63-06fc-4adf-897d-48a25349238e.json ADDED
@@ -0,0 +1,477 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/medqa/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MedQA",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "medqa",
13
+ "benchmark_name": "MedQA",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "healthcare",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "medqa:asian:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.medqa.asian",
38
+ "source_data": {
39
+ "dataset_name": "MedQA - Asian",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/medqa"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "medqa",
46
+ "task_key": "asian",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Asian).",
53
+ "metric_id": "vals_ai.medqa.asian.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 81.7,
69
+ "details": {
70
+ "benchmark_slug": "medqa",
71
+ "benchmark_name": "MedQA",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "asian",
74
+ "task_name": "Asian",
75
+ "dataset_type": "public",
76
+ "industry": "healthcare",
77
+ "raw_score": "81.7",
78
+ "raw_stderr": "1.694",
79
+ "latency": "1.928",
80
+ "cost_per_test": "0.000115",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 1.694,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ },
98
+ {
99
+ "evaluation_result_id": "medqa:black:google/gemini-2.0-flash-001:score",
100
+ "evaluation_name": "vals_ai.medqa.black",
101
+ "source_data": {
102
+ "dataset_name": "MedQA - Black",
103
+ "source_type": "url",
104
+ "url": [
105
+ "https://www.vals.ai/benchmarks/medqa"
106
+ ],
107
+ "additional_details": {
108
+ "benchmark_slug": "medqa",
109
+ "task_key": "black",
110
+ "dataset_type": "public",
111
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
112
+ }
113
+ },
114
+ "metric_config": {
115
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Black).",
116
+ "metric_id": "vals_ai.medqa.black.accuracy",
117
+ "metric_name": "Accuracy",
118
+ "metric_kind": "accuracy",
119
+ "metric_unit": "percent",
120
+ "lower_is_better": false,
121
+ "score_type": "continuous",
122
+ "min_score": 0.0,
123
+ "max_score": 100.0,
124
+ "additional_details": {
125
+ "score_scale": "percent_0_to_100",
126
+ "max_score_source": "fixed_percentage_bound",
127
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
128
+ }
129
+ },
130
+ "score_details": {
131
+ "score": 80.45,
132
+ "details": {
133
+ "benchmark_slug": "medqa",
134
+ "benchmark_name": "MedQA",
135
+ "benchmark_updated": "2026-04-16",
136
+ "task_key": "black",
137
+ "task_name": "Black",
138
+ "dataset_type": "public",
139
+ "industry": "healthcare",
140
+ "raw_score": "80.45",
141
+ "raw_stderr": "1.737",
142
+ "latency": "1.933",
143
+ "cost_per_test": "0.000115",
144
+ "max_output_tokens": "8192",
145
+ "provider": "Google"
146
+ },
147
+ "uncertainty": {
148
+ "standard_error": {
149
+ "value": 1.737,
150
+ "method": "vals_reported"
151
+ }
152
+ }
153
+ },
154
+ "generation_config": {
155
+ "generation_args": {
156
+ "max_tokens": 8192,
157
+ "max_attempts": 1
158
+ }
159
+ }
160
+ },
161
+ {
162
+ "evaluation_result_id": "medqa:hispanic:google/gemini-2.0-flash-001:score",
163
+ "evaluation_name": "vals_ai.medqa.hispanic",
164
+ "source_data": {
165
+ "dataset_name": "MedQA - Hispanic",
166
+ "source_type": "url",
167
+ "url": [
168
+ "https://www.vals.ai/benchmarks/medqa"
169
+ ],
170
+ "additional_details": {
171
+ "benchmark_slug": "medqa",
172
+ "task_key": "hispanic",
173
+ "dataset_type": "public",
174
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
175
+ }
176
+ },
177
+ "metric_config": {
178
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Hispanic).",
179
+ "metric_id": "vals_ai.medqa.hispanic.accuracy",
180
+ "metric_name": "Accuracy",
181
+ "metric_kind": "accuracy",
182
+ "metric_unit": "percent",
183
+ "lower_is_better": false,
184
+ "score_type": "continuous",
185
+ "min_score": 0.0,
186
+ "max_score": 100.0,
187
+ "additional_details": {
188
+ "score_scale": "percent_0_to_100",
189
+ "max_score_source": "fixed_percentage_bound",
190
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
191
+ }
192
+ },
193
+ "score_details": {
194
+ "score": 81.0,
195
+ "details": {
196
+ "benchmark_slug": "medqa",
197
+ "benchmark_name": "MedQA",
198
+ "benchmark_updated": "2026-04-16",
199
+ "task_key": "hispanic",
200
+ "task_name": "Hispanic",
201
+ "dataset_type": "public",
202
+ "industry": "healthcare",
203
+ "raw_score": "81",
204
+ "raw_stderr": "1.719",
205
+ "latency": "1.979",
206
+ "cost_per_test": "0.000115",
207
+ "max_output_tokens": "8192",
208
+ "provider": "Google"
209
+ },
210
+ "uncertainty": {
211
+ "standard_error": {
212
+ "value": 1.719,
213
+ "method": "vals_reported"
214
+ }
215
+ }
216
+ },
217
+ "generation_config": {
218
+ "generation_args": {
219
+ "max_tokens": 8192,
220
+ "max_attempts": 1
221
+ }
222
+ }
223
+ },
224
+ {
225
+ "evaluation_result_id": "medqa:indigenous:google/gemini-2.0-flash-001:score",
226
+ "evaluation_name": "vals_ai.medqa.indigenous",
227
+ "source_data": {
228
+ "dataset_name": "MedQA - Indigenous",
229
+ "source_type": "url",
230
+ "url": [
231
+ "https://www.vals.ai/benchmarks/medqa"
232
+ ],
233
+ "additional_details": {
234
+ "benchmark_slug": "medqa",
235
+ "task_key": "indigenous",
236
+ "dataset_type": "public",
237
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
238
+ }
239
+ },
240
+ "metric_config": {
241
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Indigenous).",
242
+ "metric_id": "vals_ai.medqa.indigenous.accuracy",
243
+ "metric_name": "Accuracy",
244
+ "metric_kind": "accuracy",
245
+ "metric_unit": "percent",
246
+ "lower_is_better": false,
247
+ "score_type": "continuous",
248
+ "min_score": 0.0,
249
+ "max_score": 100.0,
250
+ "additional_details": {
251
+ "score_scale": "percent_0_to_100",
252
+ "max_score_source": "fixed_percentage_bound",
253
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
254
+ }
255
+ },
256
+ "score_details": {
257
+ "score": 80.8,
258
+ "details": {
259
+ "benchmark_slug": "medqa",
260
+ "benchmark_name": "MedQA",
261
+ "benchmark_updated": "2026-04-16",
262
+ "task_key": "indigenous",
263
+ "task_name": "Indigenous",
264
+ "dataset_type": "public",
265
+ "industry": "healthcare",
266
+ "raw_score": "80.8",
267
+ "raw_stderr": "1.726",
268
+ "latency": "2.007",
269
+ "cost_per_test": "0.000116",
270
+ "max_output_tokens": "8192",
271
+ "provider": "Google"
272
+ },
273
+ "uncertainty": {
274
+ "standard_error": {
275
+ "value": 1.726,
276
+ "method": "vals_reported"
277
+ }
278
+ }
279
+ },
280
+ "generation_config": {
281
+ "generation_args": {
282
+ "max_tokens": 8192,
283
+ "max_attempts": 1
284
+ }
285
+ }
286
+ },
287
+ {
288
+ "evaluation_result_id": "medqa:overall:google/gemini-2.0-flash-001:score",
289
+ "evaluation_name": "vals_ai.medqa.overall",
290
+ "source_data": {
291
+ "dataset_name": "MedQA - Overall",
292
+ "source_type": "url",
293
+ "url": [
294
+ "https://www.vals.ai/benchmarks/medqa"
295
+ ],
296
+ "additional_details": {
297
+ "benchmark_slug": "medqa",
298
+ "task_key": "overall",
299
+ "dataset_type": "public",
300
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
301
+ }
302
+ },
303
+ "metric_config": {
304
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Overall).",
305
+ "metric_id": "vals_ai.medqa.overall.accuracy",
306
+ "metric_name": "Accuracy",
307
+ "metric_kind": "accuracy",
308
+ "metric_unit": "percent",
309
+ "lower_is_better": false,
310
+ "score_type": "continuous",
311
+ "min_score": 0.0,
312
+ "max_score": 100.0,
313
+ "additional_details": {
314
+ "score_scale": "percent_0_to_100",
315
+ "max_score_source": "fixed_percentage_bound",
316
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
317
+ }
318
+ },
319
+ "score_details": {
320
+ "score": 81.467,
321
+ "details": {
322
+ "benchmark_slug": "medqa",
323
+ "benchmark_name": "MedQA",
324
+ "benchmark_updated": "2026-04-16",
325
+ "task_key": "overall",
326
+ "task_name": "Overall",
327
+ "dataset_type": "public",
328
+ "industry": "healthcare",
329
+ "raw_score": "81.467",
330
+ "raw_stderr": "0.695",
331
+ "latency": "1.941",
332
+ "cost_per_test": "0.000114",
333
+ "max_output_tokens": "8192",
334
+ "provider": "Google"
335
+ },
336
+ "uncertainty": {
337
+ "standard_error": {
338
+ "value": 0.695,
339
+ "method": "vals_reported"
340
+ }
341
+ }
342
+ },
343
+ "generation_config": {
344
+ "generation_args": {
345
+ "max_tokens": 8192,
346
+ "max_attempts": 1
347
+ }
348
+ }
349
+ },
350
+ {
351
+ "evaluation_result_id": "medqa:unbiased:google/gemini-2.0-flash-001:score",
352
+ "evaluation_name": "vals_ai.medqa.unbiased",
353
+ "source_data": {
354
+ "dataset_name": "MedQA - Unbiased",
355
+ "source_type": "url",
356
+ "url": [
357
+ "https://www.vals.ai/benchmarks/medqa"
358
+ ],
359
+ "additional_details": {
360
+ "benchmark_slug": "medqa",
361
+ "task_key": "unbiased",
362
+ "dataset_type": "public",
363
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
364
+ }
365
+ },
366
+ "metric_config": {
367
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (Unbiased).",
368
+ "metric_id": "vals_ai.medqa.unbiased.accuracy",
369
+ "metric_name": "Accuracy",
370
+ "metric_kind": "accuracy",
371
+ "metric_unit": "percent",
372
+ "lower_is_better": false,
373
+ "score_type": "continuous",
374
+ "min_score": 0.0,
375
+ "max_score": 100.0,
376
+ "additional_details": {
377
+ "score_scale": "percent_0_to_100",
378
+ "max_score_source": "fixed_percentage_bound",
379
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
380
+ }
381
+ },
382
+ "score_details": {
383
+ "score": 82.65,
384
+ "details": {
385
+ "benchmark_slug": "medqa",
386
+ "benchmark_name": "MedQA",
387
+ "benchmark_updated": "2026-04-16",
388
+ "task_key": "unbiased",
389
+ "task_name": "Unbiased",
390
+ "dataset_type": "public",
391
+ "industry": "healthcare",
392
+ "raw_score": "82.65",
393
+ "raw_stderr": "1.659",
394
+ "latency": "1.873",
395
+ "cost_per_test": "0.000109",
396
+ "max_output_tokens": "8192",
397
+ "provider": "Google"
398
+ },
399
+ "uncertainty": {
400
+ "standard_error": {
401
+ "value": 1.659,
402
+ "method": "vals_reported"
403
+ }
404
+ }
405
+ },
406
+ "generation_config": {
407
+ "generation_args": {
408
+ "max_tokens": 8192,
409
+ "max_attempts": 1
410
+ }
411
+ }
412
+ },
413
+ {
414
+ "evaluation_result_id": "medqa:white:google/gemini-2.0-flash-001:score",
415
+ "evaluation_name": "vals_ai.medqa.white",
416
+ "source_data": {
417
+ "dataset_name": "MedQA - White",
418
+ "source_type": "url",
419
+ "url": [
420
+ "https://www.vals.ai/benchmarks/medqa"
421
+ ],
422
+ "additional_details": {
423
+ "benchmark_slug": "medqa",
424
+ "task_key": "white",
425
+ "dataset_type": "public",
426
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
427
+ }
428
+ },
429
+ "metric_config": {
430
+ "evaluation_description": "Accuracy reported by Vals.ai for MedQA (White).",
431
+ "metric_id": "vals_ai.medqa.white.accuracy",
432
+ "metric_name": "Accuracy",
433
+ "metric_kind": "accuracy",
434
+ "metric_unit": "percent",
435
+ "lower_is_better": false,
436
+ "score_type": "continuous",
437
+ "min_score": 0.0,
438
+ "max_score": 100.0,
439
+ "additional_details": {
440
+ "score_scale": "percent_0_to_100",
441
+ "max_score_source": "fixed_percentage_bound",
442
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/medqa"
443
+ }
444
+ },
445
+ "score_details": {
446
+ "score": 82.2,
447
+ "details": {
448
+ "benchmark_slug": "medqa",
449
+ "benchmark_name": "MedQA",
450
+ "benchmark_updated": "2026-04-16",
451
+ "task_key": "white",
452
+ "task_name": "White",
453
+ "dataset_type": "public",
454
+ "industry": "healthcare",
455
+ "raw_score": "82.2",
456
+ "raw_stderr": "1.676",
457
+ "latency": "1.928",
458
+ "cost_per_test": "0.000113",
459
+ "max_output_tokens": "8192",
460
+ "provider": "Google"
461
+ },
462
+ "uncertainty": {
463
+ "standard_error": {
464
+ "value": 1.676,
465
+ "method": "vals_reported"
466
+ }
467
+ }
468
+ },
469
+ "generation_config": {
470
+ "generation_args": {
471
+ "max_tokens": 8192,
472
+ "max_attempts": 1
473
+ }
474
+ }
475
+ }
476
+ ]
477
+ }
data/vals_ai/google/gemini-2.0-flash-001/c87b3ec1-7a1d-4560-8645-85a49f688348.json ADDED
@@ -0,0 +1,204 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/aime/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - AIME",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "aime",
13
+ "benchmark_name": "AIME",
14
+ "benchmark_updated": "2026-04-16",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "aime:aime_2024:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.aime.aime_2024",
38
+ "source_data": {
39
+ "dataset_name": "AIME - AIME 2024",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/aime"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "aime",
46
+ "task_key": "aime_2024",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2024).",
53
+ "metric_id": "vals_ai.aime.aime_2024.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 32.5,
69
+ "details": {
70
+ "benchmark_slug": "aime",
71
+ "benchmark_name": "AIME",
72
+ "benchmark_updated": "2026-04-16",
73
+ "task_key": "aime_2024",
74
+ "task_name": "AIME 2024",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "32.5",
78
+ "raw_stderr": "0.833",
79
+ "latency": "12.103",
80
+ "cost_per_test": "0.001136",
81
+ "provider": "Google"
82
+ },
83
+ "uncertainty": {
84
+ "standard_error": {
85
+ "value": 0.833,
86
+ "method": "vals_reported"
87
+ }
88
+ }
89
+ }
90
+ },
91
+ {
92
+ "evaluation_result_id": "aime:aime_2025:google/gemini-2.0-flash-001:score",
93
+ "evaluation_name": "vals_ai.aime.aime_2025",
94
+ "source_data": {
95
+ "dataset_name": "AIME - AIME 2025",
96
+ "source_type": "url",
97
+ "url": [
98
+ "https://www.vals.ai/benchmarks/aime"
99
+ ],
100
+ "additional_details": {
101
+ "benchmark_slug": "aime",
102
+ "task_key": "aime_2025",
103
+ "dataset_type": "public",
104
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
105
+ }
106
+ },
107
+ "metric_config": {
108
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (AIME 2025).",
109
+ "metric_id": "vals_ai.aime.aime_2025.accuracy",
110
+ "metric_name": "Accuracy",
111
+ "metric_kind": "accuracy",
112
+ "metric_unit": "percent",
113
+ "lower_is_better": false,
114
+ "score_type": "continuous",
115
+ "min_score": 0.0,
116
+ "max_score": 100.0,
117
+ "additional_details": {
118
+ "score_scale": "percent_0_to_100",
119
+ "max_score_source": "fixed_percentage_bound",
120
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
121
+ }
122
+ },
123
+ "score_details": {
124
+ "score": 27.083,
125
+ "details": {
126
+ "benchmark_slug": "aime",
127
+ "benchmark_name": "AIME",
128
+ "benchmark_updated": "2026-04-16",
129
+ "task_key": "aime_2025",
130
+ "task_name": "AIME 2025",
131
+ "dataset_type": "public",
132
+ "industry": "math",
133
+ "raw_score": "27.083",
134
+ "raw_stderr": "0.984",
135
+ "latency": "10.326",
136
+ "cost_per_test": "0.000951",
137
+ "provider": "Google"
138
+ },
139
+ "uncertainty": {
140
+ "standard_error": {
141
+ "value": 0.984,
142
+ "method": "vals_reported"
143
+ }
144
+ }
145
+ }
146
+ },
147
+ {
148
+ "evaluation_result_id": "aime:overall:google/gemini-2.0-flash-001:score",
149
+ "evaluation_name": "vals_ai.aime.overall",
150
+ "source_data": {
151
+ "dataset_name": "AIME - Overall",
152
+ "source_type": "url",
153
+ "url": [
154
+ "https://www.vals.ai/benchmarks/aime"
155
+ ],
156
+ "additional_details": {
157
+ "benchmark_slug": "aime",
158
+ "task_key": "overall",
159
+ "dataset_type": "public",
160
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
161
+ }
162
+ },
163
+ "metric_config": {
164
+ "evaluation_description": "Accuracy reported by Vals.ai for AIME (Overall).",
165
+ "metric_id": "vals_ai.aime.overall.accuracy",
166
+ "metric_name": "Accuracy",
167
+ "metric_kind": "accuracy",
168
+ "metric_unit": "percent",
169
+ "lower_is_better": false,
170
+ "score_type": "continuous",
171
+ "min_score": 0.0,
172
+ "max_score": 100.0,
173
+ "additional_details": {
174
+ "score_scale": "percent_0_to_100",
175
+ "max_score_source": "fixed_percentage_bound",
176
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/aime"
177
+ }
178
+ },
179
+ "score_details": {
180
+ "score": 29.792,
181
+ "details": {
182
+ "benchmark_slug": "aime",
183
+ "benchmark_name": "AIME",
184
+ "benchmark_updated": "2026-04-16",
185
+ "task_key": "overall",
186
+ "task_name": "Overall",
187
+ "dataset_type": "public",
188
+ "industry": "math",
189
+ "raw_score": "29.792",
190
+ "raw_stderr": "0.644",
191
+ "latency": "11.214",
192
+ "cost_per_test": "0.001044",
193
+ "provider": "Google"
194
+ },
195
+ "uncertainty": {
196
+ "standard_error": {
197
+ "value": 0.644,
198
+ "method": "vals_reported"
199
+ }
200
+ }
201
+ }
202
+ }
203
+ ]
204
+ }
data/vals_ai/google/gemini-2.0-flash-001/fbc69c3d-519b-4f50-b641-b85d8c250f5a.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/math500/google_gemini-2.0-flash-001/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MATH 500",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "math500",
13
+ "benchmark_name": "MATH 500",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-001",
27
+ "id": "google/gemini-2.0-flash-001",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-001",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "math500:overall:google/gemini-2.0-flash-001:score",
37
+ "evaluation_name": "vals_ai.math500.overall",
38
+ "source_data": {
39
+ "dataset_name": "MATH 500 - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/math500"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "math500",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MATH 500 (Overall).",
53
+ "metric_id": "vals_ai.math500.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 88.0,
69
+ "details": {
70
+ "benchmark_slug": "math500",
71
+ "benchmark_name": "MATH 500",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "88",
78
+ "raw_stderr": "2.852",
79
+ "latency": "3.367",
80
+ "cost_per_test": "0.00026",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 2.852,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-2.0-flash-exp/0edd0e94-356f-467a-8eb2-d1731327eb0d.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/math500/google_gemini-2.0-flash-exp/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MATH 500",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "math500",
13
+ "benchmark_name": "MATH 500",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-exp",
27
+ "id": "google/gemini-2.0-flash-exp",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-exp",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "math500:overall:google/gemini-2.0-flash-exp:score",
37
+ "evaluation_name": "vals_ai.math500.overall",
38
+ "source_data": {
39
+ "dataset_name": "MATH 500 - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/math500"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "math500",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MATH 500 (Overall).",
53
+ "metric_id": "vals_ai.math500.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 89.0,
69
+ "details": {
70
+ "benchmark_slug": "math500",
71
+ "benchmark_name": "MATH 500",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "89",
78
+ "raw_stderr": "2.748",
79
+ "latency": "3.352",
80
+ "cost_per_test": "0.0002",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 2.748,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-2.0-flash-exp/3fd0a48d-98f8-4252-a624-a90d89e91c93.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-2.0-flash-exp/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-exp",
27
+ "id": "google/gemini-2.0-flash-exp",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-exp",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-2.0-flash-exp:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 56.092,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "56.092",
75
+ "raw_stderr": "1.419",
76
+ "latency": "7.486",
77
+ "cost_per_test": "0.000302",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.419,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-2.0-flash-exp:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 67.744,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "67.744",
135
+ "raw_stderr": "0.915",
136
+ "latency": "7.486",
137
+ "cost_per_test": "0.000302",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.915,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-2.0-flash-exp:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 79.395,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "79.395",
195
+ "raw_stderr": "1.157",
196
+ "latency": "7.486",
197
+ "cost_per_test": "0.000302",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.157,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-exp/95074e1a-b052-4a19-b965-a760dd8e3934.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/mortgage_tax/google_gemini-2.0-flash-exp/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MortgageTax",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "mortgage_tax",
13
+ "benchmark_name": "MortgageTax",
14
+ "benchmark_updated": "2026-06-10",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-exp",
27
+ "id": "google/gemini-2.0-flash-exp",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-exp",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "mortgage_tax:numerical_extraction:google/gemini-2.0-flash-exp:score",
37
+ "evaluation_name": "vals_ai.mortgage_tax.numerical_extraction",
38
+ "source_data": {
39
+ "dataset_name": "MortgageTax - Numerical Extraction",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "mortgage_tax",
43
+ "task_key": "numerical_extraction",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Numerical Extraction).",
50
+ "metric_id": "vals_ai.mortgage_tax.numerical_extraction.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 64.229,
66
+ "details": {
67
+ "benchmark_slug": "mortgage_tax",
68
+ "benchmark_name": "MortgageTax",
69
+ "benchmark_updated": "2026-06-10",
70
+ "task_key": "numerical_extraction",
71
+ "task_name": "Numerical Extraction",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "64.229",
75
+ "raw_stderr": "2.645",
76
+ "latency": "2.911",
77
+ "cost_per_test": "0.000135",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 2.645,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "mortgage_tax:overall:google/gemini-2.0-flash-exp:score",
97
+ "evaluation_name": "vals_ai.mortgage_tax.overall",
98
+ "source_data": {
99
+ "dataset_name": "MortgageTax - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "mortgage_tax",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Overall).",
110
+ "metric_id": "vals_ai.mortgage_tax.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 60.374,
126
+ "details": {
127
+ "benchmark_slug": "mortgage_tax",
128
+ "benchmark_name": "MortgageTax",
129
+ "benchmark_updated": "2026-06-10",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "60.374",
135
+ "raw_stderr": "1.902",
136
+ "latency": "2.796",
137
+ "cost_per_test": "0.000132",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 1.902,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "mortgage_tax:semantic_extraction:google/gemini-2.0-flash-exp:score",
157
+ "evaluation_name": "vals_ai.mortgage_tax.semantic_extraction",
158
+ "source_data": {
159
+ "dataset_name": "MortgageTax - Semantic Extraction",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "mortgage_tax",
163
+ "task_key": "semantic_extraction",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for MortgageTax (Semantic Extraction).",
170
+ "metric_id": "vals_ai.mortgage_tax.semantic_extraction.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/mortgage_tax"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 56.518,
186
+ "details": {
187
+ "benchmark_slug": "mortgage_tax",
188
+ "benchmark_name": "MortgageTax",
189
+ "benchmark_updated": "2026-06-10",
190
+ "task_key": "semantic_extraction",
191
+ "task_name": "Semantic Extraction",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "56.518",
195
+ "raw_stderr": "2.735",
196
+ "latency": "2.681",
197
+ "cost_per_test": "0.00013",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 2.735,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-thinking-exp-01-21/39409693-1438-4025-a6ff-41c8c695b98e.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-2.0-flash-thinking-exp-01-21/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-thinking-exp-01-21",
27
+ "id": "google/gemini-2.0-flash-thinking-exp-01-21",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-thinking-exp-01-21",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-2.0-flash-thinking-exp-01-21:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 58.299,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "58.299",
75
+ "raw_stderr": "1.41",
76
+ "latency": "12.807",
77
+ "cost_per_test": "0.000827",
78
+ "max_output_tokens": "16384",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.41,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 16384,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-2.0-flash-thinking-exp-01-21:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 69.788,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "69.788",
135
+ "raw_stderr": "0.899",
136
+ "latency": "12.807",
137
+ "cost_per_test": "0.000827",
138
+ "max_output_tokens": "16384",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.899,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 16384,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-2.0-flash-thinking-exp-01-21:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 81.276,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "81.276",
195
+ "raw_stderr": "1.116",
196
+ "latency": "12.807",
197
+ "cost_per_test": "0.000827",
198
+ "max_output_tokens": "16384",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.116,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 16384,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }
data/vals_ai/google/gemini-2.0-flash-thinking-exp-01-21/5af204ba-31b0-4966-b1d0-b0e8bc39ee46.json ADDED
@@ -0,0 +1,99 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/math500/google_gemini-2.0-flash-thinking-exp-01-21/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - MATH 500",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "math500",
13
+ "benchmark_name": "MATH 500",
14
+ "benchmark_updated": "2026-01-09",
15
+ "dataset_type": "public",
16
+ "industry": "math",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-flash-thinking-exp-01-21",
27
+ "id": "google/gemini-2.0-flash-thinking-exp-01-21",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-flash-thinking-exp-01-21",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "math500:overall:google/gemini-2.0-flash-thinking-exp-01-21:score",
37
+ "evaluation_name": "vals_ai.math500.overall",
38
+ "source_data": {
39
+ "dataset_name": "MATH 500 - Overall",
40
+ "source_type": "url",
41
+ "url": [
42
+ "https://www.vals.ai/benchmarks/math500"
43
+ ],
44
+ "additional_details": {
45
+ "benchmark_slug": "math500",
46
+ "task_key": "overall",
47
+ "dataset_type": "public",
48
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
49
+ }
50
+ },
51
+ "metric_config": {
52
+ "evaluation_description": "Accuracy reported by Vals.ai for MATH 500 (Overall).",
53
+ "metric_id": "vals_ai.math500.overall.accuracy",
54
+ "metric_name": "Accuracy",
55
+ "metric_kind": "accuracy",
56
+ "metric_unit": "percent",
57
+ "lower_is_better": false,
58
+ "score_type": "continuous",
59
+ "min_score": 0.0,
60
+ "max_score": 100.0,
61
+ "additional_details": {
62
+ "score_scale": "percent_0_to_100",
63
+ "max_score_source": "fixed_percentage_bound",
64
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/math500"
65
+ }
66
+ },
67
+ "score_details": {
68
+ "score": 84.6,
69
+ "details": {
70
+ "benchmark_slug": "math500",
71
+ "benchmark_name": "MATH 500",
72
+ "benchmark_updated": "2026-01-09",
73
+ "task_key": "overall",
74
+ "task_name": "Overall",
75
+ "dataset_type": "public",
76
+ "industry": "math",
77
+ "raw_score": "84.6",
78
+ "raw_stderr": "3.163",
79
+ "latency": "11.801",
80
+ "cost_per_test": "0.000387",
81
+ "max_output_tokens": "8192",
82
+ "provider": "Google"
83
+ },
84
+ "uncertainty": {
85
+ "standard_error": {
86
+ "value": 3.163,
87
+ "method": "vals_reported"
88
+ }
89
+ }
90
+ },
91
+ "generation_config": {
92
+ "generation_args": {
93
+ "max_tokens": 8192,
94
+ "max_attempts": 1
95
+ }
96
+ }
97
+ }
98
+ ]
99
+ }
data/vals_ai/google/gemini-2.0-pro-exp-02-05/17466b64-ee76-4942-9926-aabdbfa839f0.json ADDED
@@ -0,0 +1,216 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "schema_version": "0.2.2",
3
+ "evaluation_id": "vals-ai/tax_eval_v2/google_gemini-2.0-pro-exp-02-05/1782803555.768981",
4
+ "retrieved_timestamp": "1782803555.768981",
5
+ "source_metadata": {
6
+ "source_name": "Vals.ai Leaderboard - TaxEval v2",
7
+ "source_type": "documentation",
8
+ "source_organization_name": "Vals.ai",
9
+ "source_organization_url": "https://www.vals.ai",
10
+ "evaluator_relationship": "third_party",
11
+ "additional_details": {
12
+ "benchmark_slug": "tax_eval_v2",
13
+ "benchmark_name": "TaxEval v2",
14
+ "benchmark_updated": "2026-06-17",
15
+ "dataset_type": "private",
16
+ "industry": "finance",
17
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2",
18
+ "extraction_method": "static_astro_benchmark_view_props"
19
+ }
20
+ },
21
+ "eval_library": {
22
+ "name": "Vals.ai",
23
+ "version": "unknown"
24
+ },
25
+ "model_info": {
26
+ "name": "gemini-2.0-pro-exp-02-05",
27
+ "id": "google/gemini-2.0-pro-exp-02-05",
28
+ "developer": "google",
29
+ "additional_details": {
30
+ "vals_model_id": "google/gemini-2.0-pro-exp-02-05",
31
+ "vals_provider": "Google"
32
+ }
33
+ },
34
+ "evaluation_results": [
35
+ {
36
+ "evaluation_result_id": "tax_eval_v2:correctness:google/gemini-2.0-pro-exp-02-05:score",
37
+ "evaluation_name": "vals_ai.tax_eval_v2.correctness",
38
+ "source_data": {
39
+ "dataset_name": "TaxEval v2 - Correctness",
40
+ "source_type": "other",
41
+ "additional_details": {
42
+ "benchmark_slug": "tax_eval_v2",
43
+ "task_key": "correctness",
44
+ "dataset_type": "private",
45
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
46
+ }
47
+ },
48
+ "metric_config": {
49
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Correctness).",
50
+ "metric_id": "vals_ai.tax_eval_v2.correctness.accuracy",
51
+ "metric_name": "Accuracy",
52
+ "metric_kind": "accuracy",
53
+ "metric_unit": "percent",
54
+ "lower_is_better": false,
55
+ "score_type": "continuous",
56
+ "min_score": 0.0,
57
+ "max_score": 100.0,
58
+ "additional_details": {
59
+ "score_scale": "percent_0_to_100",
60
+ "max_score_source": "fixed_percentage_bound",
61
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
62
+ }
63
+ },
64
+ "score_details": {
65
+ "score": 57.318,
66
+ "details": {
67
+ "benchmark_slug": "tax_eval_v2",
68
+ "benchmark_name": "TaxEval v2",
69
+ "benchmark_updated": "2026-06-17",
70
+ "task_key": "correctness",
71
+ "task_name": "Correctness",
72
+ "dataset_type": "private",
73
+ "industry": "finance",
74
+ "raw_score": "57.318",
75
+ "raw_stderr": "1.414",
76
+ "latency": "8.989",
77
+ "cost_per_test": "0.004012",
78
+ "max_output_tokens": "8192",
79
+ "provider": "Google"
80
+ },
81
+ "uncertainty": {
82
+ "standard_error": {
83
+ "value": 1.414,
84
+ "method": "vals_reported"
85
+ }
86
+ }
87
+ },
88
+ "generation_config": {
89
+ "generation_args": {
90
+ "max_tokens": 8192,
91
+ "max_attempts": 1
92
+ }
93
+ }
94
+ },
95
+ {
96
+ "evaluation_result_id": "tax_eval_v2:overall:google/gemini-2.0-pro-exp-02-05:score",
97
+ "evaluation_name": "vals_ai.tax_eval_v2.overall",
98
+ "source_data": {
99
+ "dataset_name": "TaxEval v2 - Overall",
100
+ "source_type": "other",
101
+ "additional_details": {
102
+ "benchmark_slug": "tax_eval_v2",
103
+ "task_key": "overall",
104
+ "dataset_type": "private",
105
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
106
+ }
107
+ },
108
+ "metric_config": {
109
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Overall).",
110
+ "metric_id": "vals_ai.tax_eval_v2.overall.accuracy",
111
+ "metric_name": "Accuracy",
112
+ "metric_kind": "accuracy",
113
+ "metric_unit": "percent",
114
+ "lower_is_better": false,
115
+ "score_type": "continuous",
116
+ "min_score": 0.0,
117
+ "max_score": 100.0,
118
+ "additional_details": {
119
+ "score_scale": "percent_0_to_100",
120
+ "max_score_source": "fixed_percentage_bound",
121
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
122
+ }
123
+ },
124
+ "score_details": {
125
+ "score": 68.152,
126
+ "details": {
127
+ "benchmark_slug": "tax_eval_v2",
128
+ "benchmark_name": "TaxEval v2",
129
+ "benchmark_updated": "2026-06-17",
130
+ "task_key": "overall",
131
+ "task_name": "Overall",
132
+ "dataset_type": "private",
133
+ "industry": "finance",
134
+ "raw_score": "68.152",
135
+ "raw_stderr": "0.916",
136
+ "latency": "8.989",
137
+ "cost_per_test": "0.004012",
138
+ "max_output_tokens": "8192",
139
+ "provider": "Google"
140
+ },
141
+ "uncertainty": {
142
+ "standard_error": {
143
+ "value": 0.916,
144
+ "method": "vals_reported"
145
+ }
146
+ }
147
+ },
148
+ "generation_config": {
149
+ "generation_args": {
150
+ "max_tokens": 8192,
151
+ "max_attempts": 1
152
+ }
153
+ }
154
+ },
155
+ {
156
+ "evaluation_result_id": "tax_eval_v2:stepwise_reasoning:google/gemini-2.0-pro-exp-02-05:score",
157
+ "evaluation_name": "vals_ai.tax_eval_v2.stepwise_reasoning",
158
+ "source_data": {
159
+ "dataset_name": "TaxEval v2 - Stepwise Reasoning",
160
+ "source_type": "other",
161
+ "additional_details": {
162
+ "benchmark_slug": "tax_eval_v2",
163
+ "task_key": "stepwise_reasoning",
164
+ "dataset_type": "private",
165
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
166
+ }
167
+ },
168
+ "metric_config": {
169
+ "evaluation_description": "Accuracy reported by Vals.ai for TaxEval v2 (Stepwise Reasoning).",
170
+ "metric_id": "vals_ai.tax_eval_v2.stepwise_reasoning.accuracy",
171
+ "metric_name": "Accuracy",
172
+ "metric_kind": "accuracy",
173
+ "metric_unit": "percent",
174
+ "lower_is_better": false,
175
+ "score_type": "continuous",
176
+ "min_score": 0.0,
177
+ "max_score": 100.0,
178
+ "additional_details": {
179
+ "score_scale": "percent_0_to_100",
180
+ "max_score_source": "fixed_percentage_bound",
181
+ "leaderboard_page_url": "https://www.vals.ai/benchmarks/tax_eval_v2"
182
+ }
183
+ },
184
+ "score_details": {
185
+ "score": 78.986,
186
+ "details": {
187
+ "benchmark_slug": "tax_eval_v2",
188
+ "benchmark_name": "TaxEval v2",
189
+ "benchmark_updated": "2026-06-17",
190
+ "task_key": "stepwise_reasoning",
191
+ "task_name": "Stepwise Reasoning",
192
+ "dataset_type": "private",
193
+ "industry": "finance",
194
+ "raw_score": "78.986",
195
+ "raw_stderr": "1.165",
196
+ "latency": "8.989",
197
+ "cost_per_test": "0.004012",
198
+ "max_output_tokens": "8192",
199
+ "provider": "Google"
200
+ },
201
+ "uncertainty": {
202
+ "standard_error": {
203
+ "value": 1.165,
204
+ "method": "vals_reported"
205
+ }
206
+ }
207
+ },
208
+ "generation_config": {
209
+ "generation_args": {
210
+ "max_tokens": 8192,
211
+ "max_attempts": 1
212
+ }
213
+ }
214
+ }
215
+ ]
216
+ }