Mistral Large 4 Preview - Surge Human Eval on Code (1-5 Rating): leaderboard
Mistral Large 4 public-preview post (2026-10-06): blind human ratings of coding answers by Surge AI annotators on a 1-5 scale, five models; values are the post's two-decimal prose figures, which round to the chart's bars. A per-model mean rating rather than the head-to-head preference shares of the October 29 release post. Mean rating (1-5), higher is better.
Metric: Mean Annotator Rating (1-5). Source: mistral.ai. Saturation forecast: Around February 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 5 | 4.22 | #16 |
| 2 | Mistral Large 4 | 3.74 | #106 |
| 3 | GLM-5.3 | 3.6 | #44 |
| 4 | Kimi K3 | 3.59 | #35 |
| 5 | GLM-5.2 | 3.4 | #80 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=mistral-large-4-preview-surge-human-eval-on-code-1-5-rating · How It Works · Data refreshed daily, snapshot 2026-11-05.