EvoEval Creative — leaderboard
Metric: Pass@1 (%). Source: evo-eval.github.io. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 | 66 |
| 2 | GPT-4 Turbo | 61 |
| 3 | Claude 3 Haiku | 47 |
| 4 | Claude 2 | 42 |
| 5 | Gemini 1.0 Pro | 40 |
| 6 | deepseek-coder-6.7B-instruct | 37 |
| 7 | deepseek-coder-1.3B-instruct | 25 |
| 8 | Phi-2 | 23 |
| 9 | starcoder2-15B | 19 |
| 10 | Mixtral 8x7B Instruct | 18 |
| 11 | starcoder2-7B | 17 |
| 12 | Mistral 7B | 14 |
| 13 | starcoder2-3B | 14 |
| 14 | gemma-7B | 13 |
| 15 | starcoder | 11 |
Interactive version: aibenchmarks.dev/benchmark?slug=evoeval-creative · How the rankings work · Data refreshed daily, snapshot 2026-07-08.