EvoEval — leaderboard
Metric: Pass@1 (%). Source: evo-eval.github.io. 51 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 Turbo | 68.63 |
| 2 | GPT-4 | 68.57 |
| 3 | Claude 3 Haiku | 56.96 |
| 4 | deepseek-coder-6.7B-instruct | 53.24 |
| 5 | Claude 2 | 51.68 |
| 6 | Gemini 1.0 Pro | 48.81 |
| 7 | deepseek-coder-1.3B-instruct | 40.96 |
| 8 | Phi-2 | 36.6 |
| 9 | Qwen-14B | 35.55 |
| 10 | starcoder2-15B | 32.82 |
| 11 | Mixtral 8x7B Instruct | 32.6 |
| 12 | starcoder2-7B | 27.87 |
| 13 | starcoder | 26.13 |
| 14 | gemma-7B | 25.29 |
| 15 | Qwen-7B | 24.16 |
Interactive version: theaggregate.ai/benchmark?slug=evoeval · How the rankings work · Data refreshed daily, snapshot 2026-07-22.