Evals for Every Language - Language bs: leaderboard
Metric: Average Score (%). Source: huggingface.co. 71 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 77.42 |
| 2 | GPT-5 | 74.87 |
| 3 | GPT-5.2 | 74.64 |
| 4 | GPT-5.3 | 74.27 |
| 5 | Claude Opus 4.8 | 74 |
| 6 | Claude Sonnet 4.5 | 73.84 |
| 7 | Claude Sonnet 4.6 | 73.32 |
| 8 | GPT-5.1 | 73.08 |
| 9 | Gemini 2.5 Pro | 72.6 |
| 10 | GPT-5.4 | 71.53 |
| 11 | Claude Sonnet 4 | 71.48 |
| 12 | GPT-5.5 | 71.46 |
| 13 | DeepSeek V3.1 | 71.17 |
| 14 | GPT-4.1 | 71.13 |
| 15 | Mistral Medium 3.1 | 71.11 |
Interactive version: theaggregate.ai/benchmark?slug=evals-for-every-language-language-bs · How It Works · Data refreshed daily, snapshot 2026-09-05.