EvoEval Concise — leaderboard
Metric: Pass@1 (%). Source: evo-eval.github.io. 51 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 | 81.1 |
| 2 | GPT-4 Turbo | 79.27 |
| 3 | Claude 3 Haiku | 71.95 |
| 4 | deepseek-coder-6.7B-instruct | 71.95 |
| 5 | Claude 2 | 69.51 |
| 6 | deepseek-coder-1.3B-instruct | 60.98 |
| 7 | Gemini 1.0 Pro | 59.76 |
| 8 | Phi-2 | 50.61 |
| 9 | Mixtral 8x7B Instruct | 43.9 |
| 10 | Qwen-14B | 43.29 |
| 11 | starcoder2-15B | 42.07 |
| 12 | starcoder2-7B | 35.37 |
| 13 | starcoder | 33.54 |
| 14 | starcoder2-3B | 33.54 |
| 15 | gemma-7B | 31.1 |
Interactive version: theaggregate.ai/benchmark?slug=evoeval-concise · How the rankings work · Data refreshed daily, snapshot 2026-07-22.