| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| CodeAssistBench (test) | GPT-5 | Accuracy54.1 | 12 | 3mo ago | |
| SWE-Bench SWE-Agent-style workflows Verified | GPT-5 | Pass@167.2 | 12 | 3mo ago | |
| HumanEvalFix held-out 25% (eval) | bayesian_greedy | Utility Difference (Δπ)5.2 | 3 | 1mo ago | |
| OSS CVDP cid016 | Solve Rate4 | 3 | 2mo ago |