CodeRouterBench: leaderboard
Model-routing study dataset (Lance1573): mean per-task score % over the held-out 2,919-task ID test split of code tasks (MBPP-style generation, bug fixing and more) for 8 frontier models.
Metric: Mean Task Score (%). Source: huggingface.co. Saturation forecast: Around April 2028. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 43.83 | #67 |
| 2 | GPT-5.4 | 42.21 | #83 |
| 3 | Claude Sonnet 4.6 | 41.31 | #94 |
| 4 | Qwen 3 Max | 39.67 | #213 |
| 5 | GLM-5 | 38.03 | #149 |
| 6 | Qwen 3.5 Plus | 37.16 | #133 |
| 7 | Kimi K2.5 | 36.66 | #152 |
| 8 | MiniMax-M2.7 | 35.62 | #270 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=coderouterbench · How It Works · Data refreshed daily, snapshot 2026-11-05.