id stringlengths 12 12 | created timestamp[s]date 2026-01-01 00:00:00 2026-01-01 00:00:00 | topic stringclasses 15 values | task_type stringclasses 12 values | difficulty stringclasses 2 values | instruction stringlengths 176 230 | input stringclasses 1 value | output stringclasses 11 values | metadata dict | hash stringlengths 40 40 |
|---|---|---|---|---|---|---|---|---|---|
train_000200 | 2026-01-01T00:00:00 | Multi-agent coordination and role separation | agent_loop | advanced | Task: agent_loop
Topic: Multi-agent coordination and role separation
Difficulty: advanced
Language: Go
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"agent",
"evaluation"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | 42fec500a899054d53d3d0df668a63bece2b1a8b | |
train_000201 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | code | expert | Task: code
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: JavaScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "JavaScript",
"moe_experts": [
"coding",
"research"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | b25383a920b9129bec8a90531138d4a3c95dfedd | |
train_000202 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | eval | advanced | Task: eval
Topic: Code-specialized model families and scaling laws
Difficulty: advanced
Language: C#
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "C#",
"moe_experts": [
"performance",
"governance"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | d1792f97c8f3215d46e4eed92968d03f4adf6d2b | |
train_000203 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | agent_loop | expert | Task: agent_loop
Topic: Extended context, memory, and repo graphs
Difficulty: expert
Language: TypeScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "TypeScript",
"moe_experts": [
"security",
"data"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 5634594a5900bf7bdebad0884bf05041cd886f48 | |
train_000204 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | design | advanced | Task: design
Topic: Tool calling protocols and sandboxed execution
Difficulty: advanced
Language: TypeScript
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "TypeScript",
"moe_experts": [
"data",
"evaluation"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | b0af7c43b056d3ab8aa36d82a8602a3811095376 | |
train_000205 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | compare | expert | Task: compare
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: expert
Language: Bash
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"governance",
"research"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | cf17d208d36ed7f9a9a61f96ef2efdf85dfc58db | |
train_000206 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | data_pipeline | advanced | Task: data_pipeline
Topic: Tool calling protocols and sandboxed execution
Difficulty: advanced
Language: JavaScript
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "JavaScript",
"moe_experts": [
"agent",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | e62e0c878567f7f90c2ad9e4856c99f22a60e8c4 | |
train_000207 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | tool_trace | advanced | Task: tool_trace
Topic: Latency, cost, throughput optimization
Difficulty: advanced
Language: JavaScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "JavaScript",
"moe_experts": [
"governance",
"performance"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | e6d30346f2e11eacdb28bc83557985b2fd2a54b0 | |
train_000208 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | compare | advanced | Task: compare
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: advanced
Language: C#
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"performance",
"coding"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 59aa170e106b077416b118b92a9d59a54314522e | |
train_000209 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | compare | advanced | Task: compare
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: advanced
Language: Go
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"security",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 1f73e11930b66938992ff0c12d2abdf6cca9d5cf | |
train_000210 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | eval | advanced | Task: eval
Topic: Latency, cost, throughput optimization
Difficulty: advanced
Language: JavaScript
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "JavaScript",
"moe_experts": [
"agent",
"evaluation"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 5a0ea2044e41ed1c87d692e634593a166e494b11 | |
train_000211 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | explain | advanced | Task: explain
Topic: Latency, cost, throughput optimization
Difficulty: advanced
Language: TypeScript
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "TypeScript",
"moe_experts": [
"performance",
"governance"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | b4eece639bf22b1a708c358c337165064fe61262 | |
train_000212 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | eval | expert | Task: eval
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: SQL
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "SQL",
"moe_experts": [
"agent",
"performance"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | 0238c469819413189826a755bd24eb0b2c984a3c | |
train_000213 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | compare | expert | Task: compare
Topic: Latency, cost, throughput optimization
Difficulty: expert
Language: Bash
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"research",
"evaluation"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | 0d2bc26a42d22415716371e72ae4d409bdc5707f | |
train_000214 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | compare | expert | Task: compare
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: SQL
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"evaluation",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 280ca86d910435cd53b669318e39c3d34870e842 | |
train_000215 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | review | expert | Task: review
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: C#
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"governance",
"security"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 169e473b368190149ba9425fc4fa5097175a4dbd | |
train_000216 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | tool_trace | advanced | Task: tool_trace
Topic: Latency, cost, throughput optimization
Difficulty: advanced
Language: Go
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Go",
"moe_experts": [
"evaluation",
"performance"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | 47c44788dfe583927f292581df4759f329729a89 | |
train_000217 | 2026-01-01T00:00:00 | Self-grading and automated model critique | compare | advanced | Task: compare
Topic: Self-grading and automated model critique
Difficulty: advanced
Language: Java
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"research",
"evaluation"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | 65fb16b1f27693c20d4d60a09aa046556a12f26c | |
train_000218 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | explain | expert | Task: explain
Topic: Extended context, memory, and repo graphs
Difficulty: expert
Language: Python
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Python",
"moe_experts": [
"data",
"performance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 8528fffeb08c21398f901f6fe7f6fe4bca5d531a | |
train_000219 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | agent_loop | expert | Task: agent_loop
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: expert
Language: C#
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"research",
"agent"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 2c0457be206f2799793c379702a8466dce0664d0 | |
train_000220 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | explain | expert | Task: explain
Topic: Model merging, distillation, and continual learning
Difficulty: expert
Language: SQL
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"agent",
"security"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | 30941d6881b540e71768f4e567edb800ea1a64c4 | |
train_000221 | 2026-01-01T00:00:00 | Multi-agent coordination and role separation | patch_diff | advanced | Task: patch_diff
Topic: Multi-agent coordination and role separation
Difficulty: advanced
Language: TypeScript
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "TypeScript",
"moe_experts": [
"research",
"governance"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 134df6a1264e4d9e7212d83016904d9330594531 | |
train_000222 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | failure_analysis | expert | Task: failure_analysis
Topic: Tool calling protocols and sandboxed execution
Difficulty: expert
Language: Rust
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Rust",
"moe_experts": [
"evaluation",
"research"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 1caa276a90ce13e12d67b3aaf04e4d2873a0d713 | |
train_000223 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | patch_diff | advanced | Task: patch_diff
Topic: Dataset curation pipelines and quality attribution
Difficulty: advanced
Language: Rust
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "Rust",
"moe_experts": [
"research",
"performance"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | 627b96c9fe341cd9edd8565de3ad918ed5dbfbd5 | |
train_000224 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | review | expert | Task: review
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: expert
Language: Bash
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"agent",
"data"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | b20522ccbcb561ec8a43639ca41963fe8e8b2508 | |
train_000225 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | design | advanced | Task: design
Topic: Latency, cost, throughput optimization
Difficulty: advanced
Language: Bash
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"governance",
"coding"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 41515788fa36682599525fb4a59d14c070d2a893 | |
train_000226 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | explain | expert | Task: explain
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: expert
Language: TypeScript
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "TypeScript",
"moe_experts": [
"research",
"performance"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 5259f27968afe07c1d95d99ebcc37a234134f9ef | |
train_000227 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | review | advanced | Task: review
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: Go
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"governance",
"coding"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | cc41f2b517d3d56b5470e56076c00e9cb6ca9ce2 | |
train_000228 | 2026-01-01T00:00:00 | Failure engineering and resilience testing | tool_trace | advanced | Task: tool_trace
Topic: Failure engineering and resilience testing
Difficulty: advanced
Language: Rust
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Rust",
"moe_experts": [
"security",
"performance"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | c9d1f2761f6f9a39c77dcedb534a3c39e0ad31b6 | |
train_000229 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | explain | advanced | Task: explain
Topic: Extended context, memory, and repo graphs
Difficulty: advanced
Language: Bash
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"agent",
"data"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | 3b8b7ddeec7032e7aaf592da871ce1d29ea1b037 | |
train_000230 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | explain | expert | Task: explain
Topic: Extended context, memory, and repo graphs
Difficulty: expert
Language: Rust
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Rust",
"moe_experts": [
"data",
"agent"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 90a6f4b85fb3f300547bbc75cdfa8c6a18c83f4a | |
train_000231 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | patch_diff | expert | Task: patch_diff
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: expert
Language: TypeScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "TypeScript",
"moe_experts": [
"coding",
"security"
],
"cost_budget": 4,
"latency_ms_target": 50,
"audit_required": true
} | df0a2f4fdbf621bbb6268d0ed57249969278c872 | |
train_000232 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | data_pipeline | expert | Task: data_pipeline
Topic: Model merging, distillation, and continual learning
Difficulty: expert
Language: Rust
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Rust",
"moe_experts": [
"performance",
"research"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 1abee00e0ca43f56e7de5d3cceb1415c06aa56c8 | |
train_000233 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | design | advanced | Task: design
Topic: Extended context, memory, and repo graphs
Difficulty: advanced
Language: Java
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"evaluation",
"performance"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | f12875779e0ac0ee302a2391fb4b0ed547480e10 | |
train_000234 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | explain | expert | Task: explain
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: Bash
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"research",
"performance"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 1c73928bf7cbda5cee404275fa124cb5d4ba4360 | |
train_000235 | 2026-01-01T00:00:00 | Failure engineering and resilience testing | review | advanced | Task: review
Topic: Failure engineering and resilience testing
Difficulty: advanced
Language: Java
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"security",
"coding"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | e3a8dfe81a22afbcb2edb1f43a739b32a211622c | |
train_000236 | 2026-01-01T00:00:00 | Failure engineering and resilience testing | compare | advanced | Task: compare
Topic: Failure engineering and resilience testing
Difficulty: advanced
Language: SQL
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"evaluation",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | b102790df21105a5e3de024495945f1771c1eff8 | |
train_000237 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | data_pipeline | advanced | Task: data_pipeline
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: Java
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"security",
"agent"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | 1d69de07268ecb130fae928fc921c62d4e92ac62 | |
train_000238 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | design | advanced | Task: design
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: JavaScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "JavaScript",
"moe_experts": [
"governance",
"research"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 500e4d9ce3d8edc5c737859ab0892c5371c95ddf | |
train_000239 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | explain | expert | Task: explain
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: expert
Language: Rust
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Rust",
"moe_experts": [
"coding",
"evaluation"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | f2e341e1f3fe0ac77e1061395cd40f6c5171dce1 | |
train_000240 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | compare | advanced | Task: compare
Topic: Dataset curation pipelines and quality attribution
Difficulty: advanced
Language: Bash
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"evaluation",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 7a4b10821dbcb4677c513d2273a03586c831862c | |
train_000241 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | tool_trace | advanced | Task: tool_trace
Topic: Tool calling protocols and sandboxed execution
Difficulty: advanced
Language: Go
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Go",
"moe_experts": [
"performance",
"coding"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 0bd6eb60727813a06ecefc257a355b3fceda64be | |
train_000242 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | eval | advanced | Task: eval
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: advanced
Language: Go
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "Go",
"moe_experts": [
"data",
"agent"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 2077cdeda56a4b8627bbbbd354e6e2d0b73c3a70 | |
train_000243 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | tool_trace | advanced | Task: tool_trace
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: TypeScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "TypeScript",
"moe_experts": [
"agent",
"governance"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 2139c90611d733c52cadf412c509cdee1ced24f4 | |
train_000244 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | tool_trace | expert | Task: tool_trace
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: Bash
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Bash",
"moe_experts": [
"data",
"coding"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | f46ad2ee9d662cbf607db6436e6dc8a827e87216 | |
train_000245 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | code | advanced | Task: code
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: Python
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Code:
```python
def agent_loop(plan, edit, test, issue, budget=4):
cost = 0
for _ in range(budget):
p = plan(issue)
patch = edit(issue, p)
ok, _ = test(patch)
cost += 1
if ok:
return patch, cost
raise RuntimeError("Budget exhausted")
``` | {
"language": "Python",
"moe_experts": [
"data",
"security"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | a3e93e46123e280ca13c89b35ed8fc692fecbf6f | |
train_000246 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | patch_diff | advanced | Task: patch_diff
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: Python
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "Python",
"moe_experts": [
"coding",
"data"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 9cd0fd5439ef46bb9df7a1d7c58d74ca78276400 | |
train_000247 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | agent_loop | advanced | Task: agent_loop
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: advanced
Language: Python
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Python",
"moe_experts": [
"coding",
"performance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 4ee74b3f12a3d3536d0e2cd9943c9d85e57e4bed | |
train_000248 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | eval | expert | Task: eval
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: Python
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "Python",
"moe_experts": [
"data",
"performance"
],
"cost_budget": 4,
"latency_ms_target": 50,
"audit_required": true
} | 29b25f32e425cec18cb38ad032b85af06c9476b9 | |
train_000249 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | design | expert | Task: design
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: Go
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"performance",
"research"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 09007bbc4b277ddf15fd60267e782d8114525c0f | |
train_000250 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | compare | advanced | Task: compare
Topic: Tool calling protocols and sandboxed execution
Difficulty: advanced
Language: Bash
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"security",
"data"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | fab1cd3119d3a15da09245dce7b835e863239429 | |
train_000251 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | explain | advanced | Task: explain
Topic: Code-specialized model families and scaling laws
Difficulty: advanced
Language: Go
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"evaluation",
"performance"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | a640e72f0f3ae0c0e367461711c1d93546037eb1 | |
train_000252 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | compare | advanced | Task: compare
Topic: Dataset curation pipelines and quality attribution
Difficulty: advanced
Language: JavaScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "JavaScript",
"moe_experts": [
"data",
"agent"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 93652f4c50c3c67418627ff5a46695120c758837 | |
train_000253 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | review | advanced | Task: review
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: advanced
Language: SQL
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"data",
"governance"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 13995636745d6bf61abfaf134aea8df17f2ea1d0 | |
train_000254 | 2026-01-01T00:00:00 | Failure engineering and resilience testing | code | advanced | Task: code
Topic: Failure engineering and resilience testing
Difficulty: advanced
Language: Bash
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"security",
"performance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 050f9ddfe46d7c31ab2fb14f525ad7860bfb957c | |
train_000255 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | design | expert | Task: design
Topic: Model merging, distillation, and continual learning
Difficulty: expert
Language: Rust
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Rust",
"moe_experts": [
"evaluation",
"coding"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 42b56960952354a610a38b5712b25321b873a392 | |
train_000256 | 2026-01-01T00:00:00 | Extended context, memory, and repo graphs | data_pipeline | expert | Task: data_pipeline
Topic: Extended context, memory, and repo graphs
Difficulty: expert
Language: SQL
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"governance",
"security"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 7bc78deedde9ee1b92f547ede8c06ee5b11f6de2 | |
train_000257 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | patch_diff | advanced | Task: patch_diff
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: advanced
Language: JavaScript
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "JavaScript",
"moe_experts": [
"research",
"governance"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 8b1e89d1eb4a57baf5a56903d58270a461a13348 | |
train_000258 | 2026-01-01T00:00:00 | Latency, cost, throughput optimization | failure_analysis | expert | Task: failure_analysis
Topic: Latency, cost, throughput optimization
Difficulty: expert
Language: Java
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Java",
"moe_experts": [
"security",
"evaluation"
],
"cost_budget": 4,
"latency_ms_target": 50,
"audit_required": true
} | 3cffa656bbc8560ad00b210a6f8f8a5a87ac472b | |
train_000259 | 2026-01-01T00:00:00 | Self-grading and automated model critique | compare | advanced | Task: compare
Topic: Self-grading and automated model critique
Difficulty: advanced
Language: TypeScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "TypeScript",
"moe_experts": [
"agent",
"evaluation"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 89c9233d449e27829890261f5c5f6f0475bd12a9 | |
train_000260 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | review | advanced | Task: review
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: advanced
Language: SQL
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"evaluation",
"security"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | c42e3a0348f684fdc2504fac8a0f199d766a2377 | |
train_000261 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | compare | expert | Task: compare
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: Java
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"research",
"evaluation"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 6f2be56184e62acf441771e72381fbb58dcdf263 | |
train_000262 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | failure_analysis | expert | Task: failure_analysis
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: expert
Language: Python
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Python",
"moe_experts": [
"research",
"agent"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | c8cbb6bb53a7e5a68f3ca439e97bfa4f7da57955 | |
train_000263 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | patch_diff | advanced | Task: patch_diff
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: advanced
Language: JavaScript
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "JavaScript",
"moe_experts": [
"evaluation",
"research"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 09f99163b03ed78db592008b9c57d789cd519fc2 | |
train_000264 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | compare | advanced | Task: compare
Topic: Model merging, distillation, and continual learning
Difficulty: advanced
Language: SQL
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"governance",
"performance"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 4abdbd5eaed653dde60ccd8a5de1bce2a188f706 | |
train_000265 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | patch_diff | advanced | Task: patch_diff
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: SQL
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "SQL",
"moe_experts": [
"security",
"coding"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | 43703d8ad330110881c3a69e4d9cd1ee9a0faae0 | |
train_000266 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | agent_loop | expert | Task: agent_loop
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: expert
Language: Python
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Python",
"moe_experts": [
"research",
"governance"
],
"cost_budget": 4,
"latency_ms_target": 50,
"audit_required": true
} | ec558e079b7a6b076da5a89ad2a2a654398a95f0 | |
train_000267 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | compare | expert | Task: compare
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: Java
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"coding",
"data"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | af07e2e0f03b4c69c34dcb6e673e861e83a99db2 | |
train_000268 | 2026-01-01T00:00:00 | Self-grading and automated model critique | design | expert | Task: design
Topic: Self-grading and automated model critique
Difficulty: expert
Language: Go
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"coding",
"governance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 602b305216cb0024700a32c0ab1bcc85479de2aa | |
train_000269 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | eval | expert | Task: eval
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: C#
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "C#",
"moe_experts": [
"agent",
"governance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 58a6ac16bb85610c0fa4ab1a3173d72c3ff3c8c7 | |
train_000270 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | eval | advanced | Task: eval
Topic: Code-specialized model families and scaling laws
Difficulty: advanced
Language: SQL
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "SQL",
"moe_experts": [
"performance",
"security"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 69e41ef4030cb0463b9daaf039990dc3e5c38254 | |
train_000271 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | patch_diff | expert | Task: patch_diff
Topic: Model merging, distillation, and continual learning
Difficulty: expert
Language: Java
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "Java",
"moe_experts": [
"data",
"governance"
],
"cost_budget": 4,
"latency_ms_target": 50,
"audit_required": true
} | 0edf02213cb2f8e6382ef7ca720ee722a094c968 | |
train_000272 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | data_pipeline | expert | Task: data_pipeline
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: expert
Language: SQL
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "SQL",
"moe_experts": [
"security",
"governance"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | 8410d6a036765298d7ff7d6e32feed8f22f6e0ea | |
train_000273 | 2026-01-01T00:00:00 | Model merging, distillation, and continual learning | failure_analysis | expert | Task: failure_analysis
Topic: Model merging, distillation, and continual learning
Difficulty: expert
Language: Java
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Java",
"moe_experts": [
"research",
"agent"
],
"cost_budget": 3,
"latency_ms_target": 50,
"audit_required": true
} | e83e324e95c726a9ef5d48854ddeb740c4ea7c88 | |
train_000274 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | code | advanced | Task: code
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: advanced
Language: Java
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"evaluation",
"performance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 6d0f6f8604596d58fc1658f7df5839e93673a425 | |
train_000275 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | agent_loop | advanced | Task: agent_loop
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: advanced
Language: Java
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Java",
"moe_experts": [
"evaluation",
"coding"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 759c3ad48683ed52d92d842d7acd42efbc59bc0c | |
train_000276 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | code | advanced | Task: code
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: advanced
Language: C#
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"data",
"research"
],
"cost_budget": 2,
"latency_ms_target": 50,
"audit_required": true
} | 8dbd9a251137edfc333b6b0303f1fc884ff485f5 | |
train_000277 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | code | expert | Task: code
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: expert
Language: Go
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Go",
"moe_experts": [
"governance",
"data"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | a300cdbc9f07fd841494f1fef8f539073c318d15 | |
train_000278 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | agent_loop | advanced | Task: agent_loop
Topic: Tool calling protocols and sandboxed execution
Difficulty: advanced
Language: C#
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"security",
"evaluation"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | fb5eceaeeb86ce3df4c82c712251052582f19e3a | |
train_000279 | 2026-01-01T00:00:00 | SWE-bench real-repo evaluation and leakage control | failure_analysis | expert | Task: failure_analysis
Topic: SWE-bench real-repo evaluation and leakage control
Difficulty: expert
Language: SQL
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "SQL",
"moe_experts": [
"performance",
"research"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | 6b0b5b8e17b73a122e1f5f19fd8053169c2d6629 | |
train_000280 | 2026-01-01T00:00:00 | Tool calling protocols and sandboxed execution | failure_analysis | expert | Task: failure_analysis
Topic: Tool calling protocols and sandboxed execution
Difficulty: expert
Language: C#
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "C#",
"moe_experts": [
"performance",
"research"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | 4bdc2dc6ee8f86f7551b09822ba06cbd608f3202 | |
train_000281 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | explain | expert | Task: explain
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: expert
Language: Bash
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"data",
"agent"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 8eb7697a47afd91657cb1b85a8730ee07c32cd10 | |
train_000282 | 2026-01-01T00:00:00 | Self-grading and automated model critique | eval | advanced | Task: eval
Topic: Self-grading and automated model critique
Difficulty: advanced
Language: SQL
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "SQL",
"moe_experts": [
"governance",
"evaluation"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | 6fb466e271b43ff7ba859bc07a4cc2b5e6e15e2b | |
train_000283 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | design | expert | Task: design
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: expert
Language: Rust
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Rust",
"moe_experts": [
"research",
"governance"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 791e7f72f39fddee7bd68a28223ecb15cf25cf79 | |
train_000284 | 2026-01-01T00:00:00 | Secure code generation, red-teaming, and policy gates | review | expert | Task: review
Topic: Secure code generation, red-teaming, and policy gates
Difficulty: expert
Language: JavaScript
Context: Offline air-gapped deployment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "JavaScript",
"moe_experts": [
"governance",
"agent"
],
"cost_budget": 3,
"latency_ms_target": 200,
"audit_required": true
} | f8296a93eb3aeda83a9504f0cf1619e37f9820a8 | |
train_000285 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | code | advanced | Task: code
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: Python
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Code:
```python
def agent_loop(plan, edit, test, issue, budget=4):
cost = 0
for _ in range(budget):
p = plan(issue)
patch = edit(issue, p)
ok, _ = test(patch)
cost += 1
if ok:
return patch, cost
raise RuntimeError("Budget exhausted")
``` | {
"language": "Python",
"moe_experts": [
"security",
"performance"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 800844eac4a1740202ccb7209982142d3a9d9c24 | |
train_000286 | 2026-01-01T00:00:00 | Self-grading and automated model critique | failure_analysis | expert | Task: failure_analysis
Topic: Self-grading and automated model critique
Difficulty: expert
Language: JavaScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "JavaScript",
"moe_experts": [
"coding",
"research"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | ac432b862381818844e6e7b9eff71ce7d2c45262 | |
train_000287 | 2026-01-01T00:00:00 | Dataset curation pipelines and quality attribution | patch_diff | expert | Task: patch_diff
Topic: Dataset curation pipelines and quality attribution
Difficulty: expert
Language: TypeScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "TypeScript",
"moe_experts": [
"performance",
"security"
],
"cost_budget": 4,
"latency_ms_target": 200,
"audit_required": true
} | 4ff564f542c1b70126a29779b6e487b072a40e66 | |
train_000288 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | review | expert | Task: review
Topic: Code-specialized model families and scaling laws
Difficulty: expert
Language: Bash
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "Bash",
"moe_experts": [
"research",
"coding"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 24dffe10a05c6543130960b058039a251174b88a | |
train_000289 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | review | expert | Task: review
Topic: Governance, provenance, and audit logging
Difficulty: expert
Language: C#
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Expert guidance with risks, metrics, and governance controls. | {
"language": "C#",
"moe_experts": [
"coding",
"evaluation"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | a4a7570ee82c6c366e67fa16746605988c11adca | |
train_000290 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | patch_diff | advanced | Task: patch_diff
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: SQL
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Patch:
```diff
- result = a / b
+ if b == 0:
+ raise ValueError('division by zero')
+ result = a / b
``` | {
"language": "SQL",
"moe_experts": [
"performance",
"security"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 3dd4ea844313bd5cec0ad9040232ee826dcd1c59 | |
train_000291 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | failure_analysis | expert | Task: failure_analysis
Topic: Code-specialized model families and scaling laws
Difficulty: expert
Language: Java
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Java",
"moe_experts": [
"data",
"evaluation"
],
"cost_budget": 5,
"latency_ms_target": 100,
"audit_required": true
} | 52420326ee8c1956d606134ba01aebcc0de84769 | |
train_000292 | 2026-01-01T00:00:00 | Governance, provenance, and audit logging | self_grade | advanced | Task: self_grade
Topic: Governance, provenance, and audit logging
Difficulty: advanced
Language: Java
Context: Regulated enterprise with audit requirements
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
{
"score": 7,
"confidence": 0.7,
"notes": "Passed tests; minimal diff; no security flags."
} | {
"language": "Java",
"moe_experts": [
"governance",
"data"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | dcc8be684a13e2728920c68b392df1cfc715b95d | |
train_000293 | 2026-01-01T00:00:00 | Multi-agent coordination and role separation | failure_analysis | expert | Task: failure_analysis
Topic: Multi-agent coordination and role separation
Difficulty: expert
Language: Bash
Context: Research benchmark validation environment
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "Bash",
"moe_experts": [
"governance",
"research"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 7689f86a3df1d28f7237955a68208eeffefc0840 | |
train_000294 | 2026-01-01T00:00:00 | Mixture-of-Experts (MoE) routing and sparsity | tool_trace | advanced | Task: tool_trace
Topic: Mixture-of-Experts (MoE) routing and sparsity
Difficulty: advanced
Language: JavaScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "JavaScript",
"moe_experts": [
"performance",
"governance"
],
"cost_budget": 5,
"latency_ms_target": 50,
"audit_required": true
} | 6d1b3028ede7016b5a1ad304adbb7b72a3f0e083 | |
train_000295 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | tool_trace | advanced | Task: tool_trace
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: Python
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Python",
"moe_experts": [
"performance",
"agent"
],
"cost_budget": 2,
"latency_ms_target": 200,
"audit_required": true
} | a42b138983a384ea4ef9327b4b687732eb2b9440 | |
train_000296 | 2026-01-01T00:00:00 | Reasoning-first coding models and tunable deliberation | self_grade | advanced | Task: self_grade
Topic: Reasoning-first coding models and tunable deliberation
Difficulty: advanced
Language: Python
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
{
"score": 8,
"confidence": 0.8,
"notes": "Passed tests; minimal diff; no security flags."
} | {
"language": "Python",
"moe_experts": [
"security",
"agent"
],
"cost_budget": 5,
"latency_ms_target": 200,
"audit_required": true
} | 437cbd4386bc0fc77f4206c3b32a613889b60497 | |
train_000297 | 2026-01-01T00:00:00 | Self-grading and automated model critique | eval | advanced | Task: eval
Topic: Self-grading and automated model critique
Difficulty: advanced
Language: JavaScript
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Metrics:
- pass@k
- time-to-green
- cost-per-fix
- regression rate
| {
"language": "JavaScript",
"moe_experts": [
"coding",
"performance"
],
"cost_budget": 3,
"latency_ms_target": 100,
"audit_required": true
} | 75084af31552f3a8e12cca27f803fcf6a23422b0 | |
train_000298 | 2026-01-01T00:00:00 | Agentic coding systems (plan→edit→test→reflect) | tool_trace | expert | Task: tool_trace
Topic: Agentic coding systems (plan→edit→test→reflect)
Difficulty: expert
Language: Java
Context: Latency-critical production API
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
[
{
"tool": "repo_checkout",
"args": {
"commit": "abc123"
}
},
{
"tool": "run_tests",
"args": {
"suite": "unit"
}
},
{
"tool": "apply_patch",
"args": {
"files": 1
}
},
{
"tool": "run_tests",
"args": {
"suite": "full"
}
}
] | {
"language": "Java",
"moe_experts": [
"coding",
"evaluation"
],
"cost_budget": 4,
"latency_ms_target": 100,
"audit_required": true
} | fc60df4ba04f526d8e2dadd22d5fc8161c6645fb | |
train_000299 | 2026-01-01T00:00:00 | Code-specialized model families and scaling laws | failure_analysis | expert | Task: failure_analysis
Topic: Code-specialized model families and scaling laws
Difficulty: expert
Language: C#
Context: Large monorepo with flaky tests and strict CI
Produce frontier-grade, production-valid output. | Facts:
- Modern AI coding success is measured by test pass rates, iteration cost, and reliability.
- Agentic and multi-agent systems outperform single-pass generation.
Failure: regression detected
Cause: missing edge case
Fix: add guard + test
| {
"language": "C#",
"moe_experts": [
"agent",
"performance"
],
"cost_budget": 2,
"latency_ms_target": 100,
"audit_required": true
} | 7867592e7130cf42b04102f969e97074c4295cbd |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.