ENAMEL: leaderboard
Efficiency-aware code-generation benchmark built from HumanEval problems with expert efficient reference solutions and strong test generators, reporting eff@1 alongside pass@1.
Metric: eff@1 (self-reported). Source: benchmarklist.com. Status: saturated. 32 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-4 Turbo (Preview) | 47 |
| 2 | GPT-4 (0613) | 45.4 |
| 3 | Llama 3 70B Instruct | 42.1 |
| 4 | Mixtral 8x22B Instruct | 40.8 |
| 5 | Claude 3 Opus | 40.1 |
| 6 | Claude 3 Haiku | 38.6 |
| 7 | Claude 3 Sonnet | 34.5 |
| 8 | Llama 3 8B Instruct | 34.4 |
| 9 | Mixtral 8x7B Instruct | 26.6 |
| 10 | starcoder | 19.5 |
| 11 | Mistral 7B | 15.2 |
| 12 | vicuna-13B | 12.3 |
| 13 | Vicuna-7B | 6.1 |
Interactive version: theaggregate.ai/benchmark?slug=enamel · How It Works · Data refreshed daily, snapshot 2026-09-05.