EvoCode-Bench - Task Completion: leaderboard
Metric: Full-task completion (Comp, %): share of the 26 tasks in which at least one of four attempts passes the cumulative verifier at every round through the final one (fail-stop), 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 February 2027. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 (High) | 42.3 |
| 2 | GPT-5.5 (High) | 38.5 |
| 3 | Claude Opus 4.6 | 34.6 |
| 4 | Kimi K2.6 | 23.1 |
| 5 | DeepSeek V4 Pro (High) | 19.2 |
| 6 | GLM-5.1 | 15.4 |
| 7 | Qwen 3.6 Plus (Thinking) | 15.4 |
| 8 | Gemini 3.1 Pro (Preview) (High) | 11.5 |
| 9 | Qwen 3.5 397B A17B | 0 |
| 10 | MiniMax-M2.7 | 0 |
| 11 | Seed 2.0 Pro (High) | 0 |
| 12 | DeepSeek V4 Flash (High) | 0 |
Interactive version: theaggregate.ai/benchmark?slug=evocode-bench-task-completion · How It Works · Data refreshed daily, snapshot 2026-10-07.