EvoCode-Bench - Single-Round: leaderboard
Metric: Single-round pass rate (SR, %): share of the 227 rounds solved in one attempt after Harbor fast-forwards the workspace to the reference solution of all earlier rounds, over the 26 stateful multi-round coding tasks of EvoCode-Bench (227 rounds, 5 to 15 per task, requirements that extend, correct or contradict earlier ones, cumulative executable verifiers), each model driving the Terminus-2 agent harness in a Harbor Docker workspace; higher is better. Source: arxiv.org. Saturation forecast: Around January 2027. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 | 78.9 |
| 2 | Claude Opus 4.7 (High) | 76.7 |
| 3 | GPT-5.5 (High) | 74.4 |
| 4 | GLM-5.1 | 63.9 |
| 5 | Kimi K2.6 | 59 |
| 6 | Qwen 3.6 Plus (Thinking) | 57.3 |
| 7 | DeepSeek V4 Pro (High) | 56.4 |
| 8 | Gemini 3.1 Pro (Preview) (High) | 46.7 |
| 9 | DeepSeek V4 Flash (High) | 46.3 |
| 10 | Qwen 3.5 397B A17B | 44.1 |
| 11 | MiniMax-M2.7 | 30 |
| 12 | Seed 2.0 Pro (High) | 23.8 |
Interactive version: theaggregate.ai/benchmark?slug=evocode-bench-single-round · How It Works · Data refreshed daily, snapshot 2026-10-07.