item_id stringclasses 13
values | sample_index int64 0 7 | agent_name stringclasses 2
values | model_name stringclasses 3
values | human_commit stringclasses 12
values | parent_commit stringclasses 12
values | benchmark_mode stringclasses 3
values | perf_command stringclasses 11
values | llm_model stringclasses 7
values | status stringclasses 2
values | error stringclasses 2
values | benchmark_type stringclasses 3
values | duration_s float64 5.69 131 | ttft_mean_ms float64 200 1.06k ⌀ | ttft_median_ms float64 210 1.11k ⌀ | ttft_p99_ms float64 289 1.32k ⌀ | tpot_mean_ms float64 14.8 41.2 ⌀ | tpot_median_ms float64 9.78 39.7 ⌀ | tpot_p99_ms float64 32.1 267 ⌀ | itl_mean_ms float64 10 52 ⌀ | itl_median_ms float64 8.27 31.8 ⌀ | itl_p99_ms float64 69.7 862 ⌀ | request_throughput_req_s float64 30 117 ⌀ | output_token_throughput_tok_s float64 1.88k 7.02k ⌀ | total_token_throughput_tok_s float64 9.57k 37.1k ⌀ | latency_avg_ms float64 216 2.39k ⌀ | latency_p50_ms float64 | latency_p99_ms float64 | throughput_tok_s float64 1.18k 8.19k ⌀ | elapsed_time_s float64 3.73 3.8 ⌀ | input_throughput_tok_s float64 17.3k 17.6k ⌀ | timestamp stringdate 2026-03-27 14:05:13 2026-03-29 05:10:55 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
vllm_core-0000 | 0 | claude_code | sonnet | 015069b01741e9ecb9e604c7fe87fbdfc306ebe5 | fbefc8a78d22b20eac042c586805c7dcbfc66b1c | serving | python benchmarks/benchmark_serving.py --model Qwen/Qwen2.5-7B-Instruct --dataset-name sharegpt --request-rate 1 | Qwen/Qwen2.5-7B-Instruct | success | null | serving | 62.311631 | 606.65 | 653.84 | 988.85 | 31.12 | 20.58 | 255.13 | 21.52 | 15.07 | 351.1 | 50.79 | 3,164.59 | 16,166.26 | null | null | null | null | null | null | 2026-03-27T14:05:13.529255+00:00 |
vllm_core-0000 | 1 | claude_code | sonnet | 015069b01741e9ecb9e604c7fe87fbdfc306ebe5 | fbefc8a78d22b20eac042c586805c7dcbfc66b1c | serving | python benchmarks/benchmark_serving.py --model Qwen/Qwen2.5-7B-Instruct --dataset-name sharegpt --request-rate 1 | Qwen/Qwen2.5-7B-Instruct | success | null | serving | 62.685172 | 596.48 | 635.58 | 980 | 31.24 | 20.69 | 260.11 | 21.49 | 15.05 | 322.71 | 51.11 | 3,184.97 | 16,270.36 | null | null | null | null | null | null | 2026-03-27T14:06:16.214923+00:00 |
vllm_core-0000 | 2 | claude_code | sonnet | 015069b01741e9ecb9e604c7fe87fbdfc306ebe5 | fbefc8a78d22b20eac042c586805c7dcbfc66b1c | serving | python benchmarks/benchmark_serving.py --model Qwen/Qwen2.5-7B-Instruct --dataset-name sharegpt --request-rate 1 | Qwen/Qwen2.5-7B-Instruct | success | null | serving | 62.377826 | 598.42 | 629.72 | 980.06 | 31.26 | 20.85 | 262.54 | 21.51 | 15.09 | 236.74 | 51.03 | 3,179.56 | 16,242.77 | null | null | null | null | null | null | 2026-03-27T14:07:18.593236+00:00 |
vllm_core-0000 | 3 | claude_code | sonnet | 015069b01741e9ecb9e604c7fe87fbdfc306ebe5 | fbefc8a78d22b20eac042c586805c7dcbfc66b1c | serving | python benchmarks/benchmark_serving.py --model Qwen/Qwen2.5-7B-Instruct --dataset-name sharegpt --request-rate 1 | Qwen/Qwen2.5-7B-Instruct | success | null | serving | 62.51957 | 609.08 | 632.82 | 992.05 | 31.38 | 20.94 | 267.18 | 21.48 | 15.1 | 246.43 | 50.79 | 3,164.57 | 16,166.18 | null | null | null | null | null | null | 2026-03-27T14:08:21.113271+00:00 |
vllm_core-0000 | 0 | codex_cli | gpt-5 | 015069b01741e9ecb9e604c7fe87fbdfc306ebe5 | fbefc8a78d22b20eac042c586805c7dcbfc66b1c | serving | python benchmarks/benchmark_serving.py --model Qwen/Qwen2.5-7B-Instruct --dataset-name sharegpt --request-rate 1 | Qwen/Qwen2.5-7B-Instruct | success | null | serving | 62.336511 | 595.09 | 625.41 | 977.73 | 31.24 | 20.82 | 263.83 | 21.47 | 15.07 | 235.8 | 51.17 | 3,188.31 | 16,287.47 | null | null | null | null | null | null | 2026-03-27T14:09:23.450705+00:00 |
vllm_core-0009 | 5 | claude_code | sonnet | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 61.685717 | 213.44 | 220.48 | 300.11 | 14.76 | 10.06 | 141.26 | 10.22 | 8.35 | 69.7 | 114.84 | 6,868.37 | 36,266.51 | null | null | null | null | null | null | 2026-03-27T14:21:20.423590+00:00 |
vllm_core-0009 | 7 | claude_code | sonnet | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.699862 | 203.42 | 215.73 | 289.37 | 14.76 | 9.78 | 165.35 | 10.04 | 8.27 | 92.67 | 117.39 | 7,021.1 | 37,072.95 | null | null | null | null | null | null | 2026-03-27T14:22:16.124177+00:00 |
vllm_core-0011 | 5 | claude_code | sonnet | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 35.977408 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,393.16 | 3.772892 | 17,392.96 | 2026-03-27T14:29:12.493336+00:00 |
vllm_core-0011 | 6 | claude_code | sonnet | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 19.935694 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,352.18 | 3.80236 | 17,260.79 | 2026-03-27T14:29:32.429535+00:00 |
vllm_core-0013 | 1 | claude_code | sonnet | 30172b4947c52890b808c6da3a6c7580f55cbb74 | a4d577b37944cbfa1bc62e4869667d1e2739d62a | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 73.839993 | 627.66 | 689.79 | 1,039.01 | 23.93 | 22.93 | 32.1 | 23.93 | 17.09 | 328.28 | 46.29 | 2,933.71 | 14,785.22 | null | null | null | null | null | null | 2026-03-27T14:36:17.833835+00:00 |
vllm_core-0001 | 2 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 5.69274 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:30:54.916159+00:00 |
vllm_core-0001 | 3 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 5.93803 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:31:00.855349+00:00 |
vllm_core-0001 | 4 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 6.052706 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:31:06.909257+00:00 |
vllm_core-0001 | 5 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 5.916767 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:31:12.827253+00:00 |
vllm_core-0001 | 6 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 6.018822 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:31:18.847278+00:00 |
vllm_core-0001 | 7 | claude_code | claude_model-claude-sonnet-4-5 | fa63e710c7fbaae3a445f669d3b5ba6b9a4ef412 | 2a0309a646b1ed83a0c40974e08c8dc628726d3c | standalone | VLLM_USE_V1=1 python3 benchmarks/benchmark_latency.py --model "/data/users/ktong/llama/llm_8b_oss" --tensor-parallel-size 1 --input_len 1000 --batch_size 32 | meta-llama/Meta-Llama-3-8B | benchmark_failed | No latency metrics found in output | latency | 5.866601 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:31:24.715097+00:00 |
vllm_core-0003 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 106.672684 | 1,056.73 | 1,114.2 | 1,303.32 | 36.28 | 34.73 | 57.97 | 35.77 | 31.75 | 230.63 | 30.04 | 1,881.86 | 9,572.72 | null | null | null | null | null | null | 2026-03-29T03:36:17.908683+00:00 | |
vllm_core-0003 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 60.971578 | 641.35 | 699.26 | 885.4 | 36.22 | 34.67 | 57.93 | 35.71 | 31.64 | 230.2 | 34.37 | 2,153.02 | 10,952.08 | null | null | null | null | null | null | 2026-03-29T03:37:18.880737+00:00 | |
vllm_core-0003 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 61.170527 | 631.5 | 689.59 | 872.72 | 36.16 | 34.6 | 57.64 | 35.65 | 31.58 | 226.01 | 34.57 | 2,161.11 | 11,011.58 | null | null | null | null | null | null | 2026-03-29T03:38:20.051823+00:00 | |
vllm_core-0003 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | benchmark_failed | vLLM server crashed after applying patch | serving | 8.977424 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 2026-03-29T03:38:29.029699+00:00 | |
vllm_core-0003 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 61.40068 | 650.73 | 694.22 | 888.14 | 35.72 | 34.92 | 42.39 | 35.66 | 31.76 | 124.94 | 34.27 | 2,179.4 | 10,951.85 | null | null | null | null | null | null | 2026-03-29T03:39:30.430861+00:00 | |
vllm_core-0003 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 61.263455 | 643.79 | 694.82 | 883.33 | 35.79 | 34.79 | 44.95 | 35.67 | 31.71 | 119.92 | 34.38 | 2,172.61 | 10,974.47 | null | null | null | null | null | null | 2026-03-29T03:40:31.694847+00:00 | |
vllm_core-0003 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 60.84568 | 643.41 | 701.5 | 887.9 | 35.91 | 34.71 | 44.96 | 35.72 | 31.67 | 205.72 | 34.32 | 2,169.48 | 10,954.46 | null | null | null | null | null | null | 2026-03-29T03:41:32.541055+00:00 | |
vllm_core-0003 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 19d98e0c7db96713f0e2201649159431177a56e2 | 2b04c209ee98174f29f1fc98f0dc3222d652a7bd | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | success | null | serving | 60.821189 | 641.8 | 699.05 | 883.5 | 36.16 | 34.62 | 57.73 | 35.66 | 31.66 | 227.82 | 34.42 | 2,156.29 | 10,968.72 | null | null | null | null | null | null | 2026-03-29T03:42:33.362780+00:00 | |
vllm_core-0004 | 0 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 75.451752 | 654.48 | 676.61 | 1,107.16 | 24.3 | 23.2 | 33.79 | 23.92 | 16.51 | 272.61 | 45.94 | 2,912.03 | 14,671.51 | null | null | null | null | null | null | 2026-03-29T03:47:04.264696+00:00 |
vllm_core-0004 | 1 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.042356 | 659.44 | 702.1 | 1,111.48 | 24.82 | 23.39 | 33.97 | 24.41 | 16.66 | 271.87 | 45.16 | 2,863.02 | 14,424.59 | null | null | null | null | null | null | 2026-03-29T03:47:59.307537+00:00 |
vllm_core-0004 | 2 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 54.564632 | 661.73 | 709.64 | 1,106.74 | 24.86 | 23.35 | 34.15 | 24.49 | 16.71 | 267.28 | 45.06 | 2,856.3 | 14,390.72 | null | null | null | null | null | null | 2026-03-29T03:48:53.872760+00:00 |
vllm_core-0004 | 3 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 54.43892 | 658.06 | 701.06 | 1,111.55 | 24.4 | 22.97 | 33.66 | 24.02 | 16.56 | 274.24 | 45.73 | 2,898.88 | 14,605.28 | null | null | null | null | null | null | 2026-03-29T03:49:48.312266+00:00 |
vllm_core-0004 | 4 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 49.540464 | 660.14 | 705.47 | 1,108.51 | 24.58 | 23.12 | 33.68 | 24.19 | 16.57 | 247.02 | 45.42 | 2,879.51 | 14,507.66 | null | null | null | null | null | null | 2026-03-29T03:50:37.853266+00:00 |
vllm_core-0004 | 5 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 54.976917 | 651.48 | 684.19 | 1,100.78 | 24.5 | 23.24 | 33.84 | 24.08 | 16.58 | 279.06 | 45.7 | 2,896.84 | 14,594.98 | null | null | null | null | null | null | 2026-03-29T03:51:32.830725+00:00 |
vllm_core-0004 | 6 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 54.709602 | 652.12 | 695.05 | 1,101.88 | 24.35 | 22.93 | 33.55 | 23.95 | 16.54 | 290.29 | 45.9 | 2,909.7 | 14,659.76 | null | null | null | null | null | null | 2026-03-29T03:52:27.540827+00:00 |
vllm_core-0004 | 7 | claude_code | claude_model-claude-sonnet-4-5 | b55ed6ef8ab0dce7fb0f79ff292dafdb4d22610c | 2f385183f35497e030ef22c9820d83b83bc4f6db | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 54.726559 | 652.45 | 696.93 | 1,102.63 | 24.29 | 22.86 | 33.47 | 23.92 | 16.64 | 247.78 | 45.97 | 2,914.18 | 14,682.36 | null | null | null | null | null | null | 2026-03-29T03:53:22.267874+00:00 |
vllm_core-0005 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.752535 | 636.93 | 686.92 | 1,068.79 | 24.35 | 23.55 | 32.94 | 24.35 | 17.21 | 259.65 | 45.57 | 2,887.88 | 14,554.24 | null | null | null | null | null | null | 2026-03-29T03:57:08.714428+00:00 |
vllm_core-0005 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.51648 | 631.22 | 570.02 | 1,064.5 | 24.4 | 25.45 | 32.47 | 24.4 | 17.2 | 381 | 45.64 | 2,892.36 | 14,576.8 | null | null | null | null | null | null | 2026-03-29T03:58:04.231424+00:00 |
vllm_core-0005 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.902548 | 637.88 | 702.58 | 1,063.03 | 24.32 | 23.28 | 32.73 | 24.32 | 17.29 | 337.73 | 45.51 | 2,884.08 | 14,535.11 | null | null | null | null | null | null | 2026-03-29T03:59:00.134511+00:00 |
vllm_core-0005 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.438446 | 633.75 | 570.41 | 1,070.57 | 24.5 | 25.58 | 32.6 | 24.5 | 17.27 | 386.56 | 45.41 | 2,877.39 | 14,501.37 | null | null | null | null | null | null | 2026-03-29T03:59:55.573496+00:00 |
vllm_core-0005 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.852084 | 643.59 | 708.37 | 1,071.55 | 24.4 | 23.36 | 32.78 | 24.4 | 17.28 | 342.73 | 45.33 | 2,872.49 | 14,476.7 | null | null | null | null | null | null | 2026-03-29T04:00:51.426113+00:00 |
vllm_core-0005 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.522285 | 638.62 | 703.67 | 1,066.73 | 24.33 | 23.29 | 32.78 | 24.33 | 17.23 | 339.84 | 45.56 | 2,887.17 | 14,550.68 | null | null | null | null | null | null | 2026-03-29T04:01:46.948914+00:00 |
vllm_core-0005 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.485052 | 635.83 | 691.45 | 1,064.75 | 24.33 | 23.45 | 32.93 | 24.33 | 17.18 | 256.59 | 45.61 | 2,890.6 | 14,567.97 | null | null | null | null | null | null | 2026-03-29T04:02:42.434475+00:00 |
vllm_core-0005 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 22d33baca2c0c639cfd45c48e99803e56c3efa74 | b0e96aaebbfbe8e70478e4192a5a13864ffdefa6 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.546632 | 637.52 | 640.05 | 1,067.36 | 24.31 | 23.11 | 32.55 | 24.31 | 17.17 | 384.89 | 45.58 | 2,888.22 | 14,555.94 | null | null | null | null | null | null | 2026-03-29T04:03:37.981671+00:00 |
vllm_core-0006 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 44.185636 | null | null | null | null | null | null | null | null | null | null | null | null | 217.032547 | null | null | 1,177.2 | null | null | 2026-03-29T04:07:24.616540+00:00 |
vllm_core-0006 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 40.869766 | null | null | null | null | null | null | null | null | null | null | null | null | 216.751162 | null | null | 1,177.2 | null | null | 2026-03-29T04:08:05.488005+00:00 |
vllm_core-0006 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.715122 | null | null | null | null | null | null | null | null | null | null | null | null | 217.899165 | null | null | 1,177.5 | null | null | 2026-03-29T04:08:45.204822+00:00 |
vllm_core-0006 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.184707 | null | null | null | null | null | null | null | null | null | null | null | null | 216.956995 | null | null | 1,177.5 | null | null | 2026-03-29T04:09:24.391300+00:00 |
vllm_core-0006 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.018949 | null | null | null | null | null | null | null | null | null | null | null | null | 216.304557 | null | null | 1,177.5 | null | null | 2026-03-29T04:10:03.412006+00:00 |
vllm_core-0006 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.930584 | null | null | null | null | null | null | null | null | null | null | null | null | 216.899741 | null | null | 1,177.5 | null | null | 2026-03-29T04:10:43.344339+00:00 |
vllm_core-0006 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.181619 | null | null | null | null | null | null | null | null | null | null | null | null | 216.209539 | null | null | 1,177.2 | null | null | 2026-03-29T04:11:22.527730+00:00 |
vllm_core-0006 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 98f47f2a4032f8c395268de80858c64ffcfc60fa | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | standalone | python benchmarks/benchmark_latency.py | unknown | success | null | latency | 39.882642 | null | null | null | null | null | null | null | null | null | null | null | null | 215.747895 | null | null | 1,177.5 | null | null | 2026-03-29T04:12:02.412130+00:00 |
vllm_core-0007 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 128.439322 | null | null | null | null | null | null | null | null | null | null | null | null | 2,371.281119 | null | null | 8,188 | null | null | 2026-03-29T04:17:12.720933+00:00 |
vllm_core-0007 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 129.322742 | null | null | null | null | null | null | null | null | null | null | null | null | 2,378.270704 | null | null | 8,161.1 | null | null | 2026-03-29T04:19:22.045709+00:00 |
vllm_core-0007 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 129.891826 | null | null | null | null | null | null | null | null | null | null | null | null | 2,376.830434 | null | null | 8,161 | null | null | 2026-03-29T04:21:31.939517+00:00 |
vllm_core-0007 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 130.846213 | null | null | null | null | null | null | null | null | null | null | null | null | 2,383.958278 | null | null | 8,159.1 | null | null | 2026-03-29T04:23:42.787788+00:00 |
vllm_core-0007 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 128.458719 | null | null | null | null | null | null | null | null | null | null | null | null | 2,385.552153 | null | null | 8,135 | null | null | 2026-03-29T04:25:51.248520+00:00 |
vllm_core-0007 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 127.099713 | null | null | null | null | null | null | null | null | null | null | null | null | 2,376.414762 | null | null | 8,172.1 | null | null | 2026-03-29T04:27:58.350254+00:00 |
vllm_core-0007 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 127.39201 | null | null | null | null | null | null | null | null | null | null | null | null | 2,376.928136 | null | null | 8,158.6 | null | null | 2026-03-29T04:30:05.744292+00:00 |
vllm_core-0007 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 8c1e77fb585c4f42783a3d88c1efc7c9e15fd89f | 5fc5ce0fe45f974fc8840175e8321652238400f0 | standalone | python benchmarks/benchmark_latency.py --model meta-llama/Meta-Llama-3-8B-Instruct --batch-size 32 --input-len 512 --output-len 128 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | latency | 128.019625 | null | null | null | null | null | null | null | null | null | null | null | null | 2,384.253927 | null | null | 8,144.9 | null | null | 2026-03-29T04:32:13.765968+00:00 |
vllm_core-0008 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.641855 | 633.39 | 634.03 | 1,067.4 | 24.41 | 23.27 | 32.63 | 24.41 | 17.19 | 391.21 | 45.67 | 2,894.3 | 14,586.6 | null | null | null | null | null | null | 2026-03-29T04:36:34.596866+00:00 |
vllm_core-0008 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 61.310124 | 642.55 | 643.63 | 1,076.95 | 24.58 | 23.45 | 32.88 | 24.58 | 17.35 | 391.37 | 45.26 | 2,868.01 | 14,454.09 | null | null | null | null | null | null | 2026-03-29T04:37:35.907503+00:00 |
vllm_core-0008 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 55.882728 | 638.78 | 688.77 | 1,071.44 | 24.39 | 23.61 | 32.93 | 24.39 | 17.16 | 261.37 | 45.57 | 2,887.62 | 14,552.92 | null | null | null | null | null | null | 2026-03-29T04:38:31.790757+00:00 |
vllm_core-0008 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 61.323037 | 647.31 | 710.08 | 1,077.54 | 24.4 | 23.4 | 32.85 | 24.4 | 17.2 | 343.57 | 45.4 | 2,876.82 | 14,498.49 | null | null | null | null | null | null | 2026-03-29T04:39:33.114305+00:00 |
vllm_core-0008 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 56.148914 | 637.03 | 689.46 | 1,066.42 | 24.34 | 23.53 | 32.88 | 24.34 | 17.13 | 259.97 | 45.65 | 2,892.57 | 14,577.9 | null | null | null | null | null | null | 2026-03-29T04:40:29.263737+00:00 |
vllm_core-0008 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 61.155091 | 632.39 | 570.16 | 1,067.85 | 24.45 | 25.48 | 32.52 | 24.45 | 17.19 | 385.01 | 45.63 | 2,891.4 | 14,572 | null | null | null | null | null | null | 2026-03-29T04:41:30.419339+00:00 |
vllm_core-0008 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 60.872657 | 634.87 | 696.7 | 1,063.9 | 24.32 | 23.34 | 32.73 | 24.32 | 17.13 | 343.67 | 45.75 | 2,899.01 | 14,610.36 | null | null | null | null | null | null | 2026-03-29T04:42:31.292544+00:00 |
vllm_core-0008 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 296f927f2493908984707354e3cc5d7b2e41650b | 0032903a5bb7c7c655f52f4efdfcc221947e9ca8 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --dtype float16 --num-prompts 300 --seed 0 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 56.210288 | 637.65 | 692.65 | 1,062.73 | 24.2 | 23.35 | 32.73 | 24.2 | 17.06 | 256.86 | 45.81 | 2,903.05 | 14,630.7 | null | null | null | null | null | null | 2026-03-29T04:43:27.503361+00:00 |
vllm_core-0009 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 61.283478 | 202.49 | 210.4 | 291.41 | 16.84 | 11.94 | 146.87 | 12.11 | 10.25 | 72.46 | 102.12 | 6,108 | 32,251.61 | null | null | null | null | null | null | 2026-03-29T04:47:34.907839+00:00 |
vllm_core-0009 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.873459 | 206.19 | 214.31 | 295.04 | 16.77 | 11.87 | 146.55 | 12.05 | 10.26 | 73.93 | 102.13 | 6,108.28 | 32,253.06 | null | null | null | null | null | null | 2026-03-29T04:48:30.781825+00:00 |
vllm_core-0009 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.775763 | 212.44 | 223.54 | 300.88 | 16.65 | 11.82 | 142.21 | 12.06 | 10.23 | 100.8 | 101.43 | 6,066.66 | 32,033.32 | null | null | null | null | null | null | 2026-03-29T04:49:26.558139+00:00 |
vllm_core-0009 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.373017 | 209.24 | 219.89 | 297.16 | 16.6 | 11.85 | 141.74 | 12.07 | 10.22 | 101.34 | 101.71 | 6,083.47 | 32,122.09 | null | null | null | null | null | null | 2026-03-29T04:50:21.931725+00:00 |
vllm_core-0009 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.870016 | 200.31 | 213.38 | 289.4 | 16.58 | 11.8 | 139.28 | 12.06 | 10.26 | 99.26 | 102.7 | 6,142.69 | 32,434.77 | null | null | null | null | null | null | 2026-03-29T04:51:17.802255+00:00 |
vllm_core-0009 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.653439 | 211.17 | 219.22 | 299.69 | 16.86 | 11.94 | 148.22 | 12.12 | 10.28 | 76.29 | 101.1 | 6,046.64 | 31,927.6 | null | null | null | null | null | null | 2026-03-29T04:52:13.456226+00:00 |
vllm_core-0009 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.352657 | 207.87 | 216.35 | 299.04 | 16.73 | 11.83 | 146.9 | 12.02 | 10.14 | 77.88 | 102.23 | 6,114.54 | 32,286.12 | null | null | null | null | null | null | 2026-03-29T04:53:08.809421+00:00 |
vllm_core-0009 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 299ebb62b269ce167eb1c71b5e39a1dc1f65ce1c | f728ab8e3578c22b42ed53e51b5e8ec35328d8b9 | serving | vllm bench serve --model Qwen/Qwen2.5-1.5B-Instruct --request-rate 1 --num-prompts 100 --random-input-len 1000 --random-output-len 100 --tokenizer Qwen/Qwen2.5-1.5B-Instruct --ignore-eos | Qwen/Qwen2.5-1.5B-Instruct | success | null | serving | 55.76768 | 204.91 | 211.11 | 293.85 | 16.88 | 11.89 | 150.12 | 12.03 | 10.21 | 75.87 | 102.45 | 6,127.51 | 32,354.58 | null | null | null | null | null | null | 2026-03-29T04:54:04.577665+00:00 |
vllm_core-0010 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 35.761963 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,451.96 | 3.73155 | 17,582.61 | 2026-03-29T04:57:31.702967+00:00 |
vllm_core-0010 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 21.562302 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,451.65 | 3.733049 | 17,581.6 | 2026-03-29T04:57:53.265741+00:00 |
vllm_core-0010 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 21.299643 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,438.41 | 3.741562 | 17,538.91 | 2026-03-29T04:58:14.565888+00:00 |
vllm_core-0010 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 19.466668 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,448.28 | 3.735562 | 17,570.72 | 2026-03-29T04:58:34.033033+00:00 |
vllm_core-0011 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 2deb029d115dadd012ce5ea70487a207cb025493 | 029c71de11bc3bcf84a1b3cf9d91e79ab6949799 | prefix_caching | python3 benchmarks/benchmark_prefix_caching.py --model RedHatAI/Meta-Llama-3-8B-Instruct-FP8 --output-len 200 --enable-prefix-caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | success | null | prefix_caching | 19.869989 | null | null | null | null | null | null | null | null | null | null | null | null | null | null | null | 5,396.87 | 3.770858 | 17,404.95 | 2026-03-29T04:59:24.024981+00:00 |
vllm_core-0012 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.37999 | 730.94 | 734.27 | 1,261.05 | 36.59 | 36.92 | 48.88 | 47.25 | 24.05 | 817.96 | 33.51 | 2,100.45 | null | null | null | null | null | null | null | 2026-03-29T05:03:08.739134+00:00 |
vllm_core-0012 | 1 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.328714 | 731 | 805.2 | 1,269.39 | 37.27 | 36.68 | 50.44 | 47.9 | 25.1 | 805.99 | 33.06 | 2,072.27 | null | null | null | null | null | null | null | 2026-03-29T05:03:47.068327+00:00 |
vllm_core-0012 | 2 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 39.038474 | 774.26 | 776.12 | 1,322.92 | 40.47 | 39.68 | 51.64 | 51.63 | 27.28 | 862.2 | 30.66 | 1,921.99 | null | null | null | null | null | null | null | 2026-03-29T05:04:26.107306+00:00 |
vllm_core-0012 | 3 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.262806 | 731.95 | 806.74 | 1,268.62 | 36.22 | 35.56 | 49.24 | 46.89 | 23.95 | 807.17 | 33.77 | 2,116.77 | null | null | null | null | null | null | null | 2026-03-29T05:05:04.370690+00:00 |
vllm_core-0012 | 4 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.320001 | 731.49 | 807.13 | 1,267.71 | 37.16 | 34.96 | 50.19 | 47.8 | 25.07 | 807.62 | 33.13 | 2,076.79 | null | null | null | null | null | null | null | 2026-03-29T05:05:42.691173+00:00 |
vllm_core-0012 | 5 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.613174 | 749.01 | 825.88 | 1,288.05 | 37.83 | 37.16 | 50.16 | 48.69 | 25.18 | 826.25 | 32.46 | 2,034.88 | null | null | null | null | null | null | null | 2026-03-29T05:06:21.304839+00:00 |
vllm_core-0012 | 6 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 38.372348 | 716.3 | 780.13 | 1,258.34 | 37.27 | 35.24 | 50.73 | 47.63 | 25.11 | 813.21 | 33.23 | 2,082.83 | null | null | null | null | null | null | null | 2026-03-29T05:06:59.677694+00:00 |
vllm_core-0012 | 7 | claude_code | claude_model-claude-sonnet-4-5 | 660470e5a36b8e52083615ad7c85e9b4fd4c72ce | 8d59dbb00044a588cab96bcdc028006ed922eb06 | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --tensor-parallel-size 1 --enable-prefix-caching --use-v2-block-manager | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 39.041483 | 747.6 | 785.96 | 1,290.22 | 41.22 | 39.54 | 54.15 | 52 | 29.84 | 842.21 | 30.42 | 1,906.78 | null | null | null | null | null | null | null | 2026-03-29T05:07:38.719699+00:00 |
vllm_core-0013 | 0 | claude_code | claude_model-claude-sonnet-4-5 | 30172b4947c52890b808c6da3a6c7580f55cbb74 | a4d577b37944cbfa1bc62e4869667d1e2739d62a | serving | python benchmarks/benchmark_serving.py --model meta-llama/Meta-Llama-3-8B-Instruct --backend vllm --num-prompts 100 | meta-llama/Meta-Llama-3-8B-Instruct | success | null | serving | 48.995219 | 640.54 | 694.38 | 1,064.94 | 24.29 | 23.42 | 32.88 | 24.29 | 17.26 | 256.68 | 45.52 | 2,884.83 | 14,538.88 | null | null | null | null | null | null | 2026-03-29T05:10:55.876316+00:00 |
Pass@k GPU Benchmark Results
GPU benchmark results for agent-generated optimization patches from ISO-Bench.
Patches sourced from Inferencebench/pass-at-k-samples, benchmarked on NVIDIA H100 80GB GPU using Docker-containerized vLLM.
Summary
- 86 total rows (79 successful benchmarks, 7 benchmark failures)
- 12 tasks benchmarked across 3 agent/model configurations
- Benchmark types: serving, standalone (latency/throughput), prefix caching
Status Breakdown
| Status | Count | Description |
|---|---|---|
success |
79 | Benchmark completed, metrics captured |
benchmark_failed |
7 | Patch applied but benchmark errored (server crash, metric parse failure) |
Agents
| Agent | Model | Successful Benchmarks |
|---|---|---|
| claude_code | claude_model-claude-sonnet-4-5 | 69 |
| claude_code | sonnet | 9 |
| codex_cli | gpt-5 | 1 |
Metrics
Serving Benchmarks
| Column | Description |
|---|---|
ttft_mean_ms |
Mean Time To First Token |
ttft_median_ms |
Median Time To First Token |
ttft_p99_ms |
P99 Time To First Token |
tpot_mean_ms |
Mean Time Per Output Token |
tpot_median_ms |
Median Time Per Output Token |
tpot_p99_ms |
P99 Time Per Output Token |
itl_mean_ms |
Mean Inter-Token Latency |
itl_median_ms |
Median Inter-Token Latency |
itl_p99_ms |
P99 Inter-Token Latency |
request_throughput_req_s |
Request throughput (req/s) |
output_token_throughput_tok_s |
Output token throughput (tok/s) |
total_token_throughput_tok_s |
Total token throughput (tok/s) |
Latency Benchmarks
| Column | Description |
|---|---|
latency_avg_ms |
Average latency (ms) |
latency_p50_ms |
P50 latency (ms) |
latency_p99_ms |
P99 latency (ms) |
throughput_tok_s |
Token throughput (tok/s) |
Prefix Caching Benchmarks
| Column | Description |
|---|---|
input_throughput_tok_s |
Input throughput (tok/s) |
throughput_tok_s |
Output throughput (tok/s) |
elapsed_time_s |
Total elapsed time (s) |
Per-Task Results
| Task | Samples | Benchmark Mode | LLM Model | Avg Throughput |
|---|---|---|---|---|
| vllm_core-0000 | 5 | serving | Qwen/Qwen2.5-7B-Instruct | 3176.4 tok/s |
| vllm_core-0003 | 7 | serving | deepseek-ai/DeepSeek-Coder-V2-Lite-Instruct | 2124.8 tok/s |
| vllm_core-0004 | 8 | serving | meta-llama/Meta-Llama-3-8B-Instruct | 2891.3 tok/s |
| vllm_core-0005 | 8 | serving | meta-llama/Meta-Llama-3-8B-Instruct | 2885.0 tok/s |
| vllm_core-0006 | 8 | standalone | unknown | 1177.4 tok/s |
| vllm_core-0007 | 8 | standalone | meta-llama/Meta-Llama-3-8B-Instruct | 8160.0 tok/s |
| vllm_core-0008 | 8 | serving | meta-llama/Meta-Llama-3-8B-Instruct | 2889.1 tok/s |
| vllm_core-0009 | 10 | serving | Qwen/Qwen2.5-1.5B-Instruct | 6268.7 tok/s |
| vllm_core-0010 | 4 | prefix_caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | 5447.6 tok/s |
| vllm_core-0011 | 3 | prefix_caching | RedHatAI/Meta-Llama-3-8B-Instruct-FP8 | 5380.7 tok/s |
| vllm_core-0012 | 8 | serving | meta-llama/Meta-Llama-3-8B-Instruct | 2039.1 tok/s |
| vllm_core-0013 | 2 | serving | meta-llama/Meta-Llama-3-8B-Instruct | 2909.3 tok/s |
Usage
from datasets import load_dataset
ds = load_dataset("Inferencebench/pass-at-k-benchmark-results", split="train")
# Filter to successful benchmarks
success = ds.filter(lambda x: x["status"] == "success")
# Get results for a specific task
task_results = success.filter(lambda x: x["item_id"] == "vllm_core-0004")
- Downloads last month
- -