HumanEval+: leaderboard
HumanEval+ code-generation leaderboard from EvalPlus.
Metric: HumanEval+ pass@1 (self-reported). Source: benchmarklist.com. Status: saturated. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O1 Mini (2024-09-12) | 89 |
| 2 | GPT-4o (2024-08-06) | 87.2 |
| 3 | Qwen 2.5 Coder 32B Instruct | 87.2 |
| 4 | GPT-4 Turbo | 86.6 |
| 5 | GPT-4o Mini (2024-07-18) | 83.5 |
| 6 | Claude 3.5 Sonnet (20240620) | 81.7 |
| 7 | Grok Beta | 80.5 |
| 8 | Gemini 1.5 Pro (002) | 79.3 |
| 9 | GPT-4 | 79.3 |
| 10 | Claude 3 Opus | 77.4 |
| 11 | OpenCoder-8B-Instruct | 77.4 |
| 12 | Gemini 1.5 Flash (002) | 75.6 |
| 13 | Codestral-22B-v0.1 | 73.8 |
| 14 | Llama 3 70B Instruct | 72 |
| 15 | Mixtral 8x22B Instruct (v0.1) | 72 |
Interactive version: theaggregate.ai/benchmark?slug=humaneval-plus · How It Works · Data refreshed daily, snapshot 2026-09-05.