HELM Arabic - Mbzuai Human Translated Arabic Mmlu - College Biology: leaderboard
Metric: Exact Match (%). Source: crfm.stanford.edu. 41 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.7 | 96.53 |
| 2 | GPT-5.4 (2026-03-05) | 93.75 |
| 3 | Gemma 4 31B (IT) | 92.36 |
| 4 | Qwen 3.5 397B A17B | 92.36 |
| 5 | Claude Sonnet 4.6 | 91.67 |
| 6 | Gemini 2.5 Flash (Thinking) | 90.97 |
| 7 | GPT-4.1 (2025-04-14) | 90.28 |
| 8 | Llama 4 Maverick Instruct FP8 | 90.28 |
| 9 | Claude Haiku 4.5 (20251001) | 88.19 |
| 10 | Mistral Large 3 | 87.5 |
| 11 | Qwen 3 Next 80B A3B Instruct | 84.72 |
| 12 | GPT-4.1 Mini | 82.64 |
| 13 | Qwen 2.5 72B Instruct | 81.94 |
| 14 | DeepSeek V3.1 | 80.56 |
| 15 | Gemini 2.5 Flash Lite (Thinking) | 79.86 |
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-mbzuai-human-translated-arabic-mmlu-college-biology · How It Works · Data refreshed daily, snapshot 2026-09-19.