HELM Arabic - Alghafa: leaderboard
The "Alghafa" subtask of HELM Arabic. Stanford HELM's Arabic project: Modern Standard Arabic knowledge, reasoning and understanding scenarios (AlGhafa, Arabic MMLU, exams and native-authored sets) run through the HELM harness with exact-match style metrics.
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Around March 2027. 41 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.7 | 79.97 | #50 |
| 2 | Claude Sonnet 4.6 | 79.72 | #92 |
| 3 | GPT-4.1 (2025-04-14) | 79.49 | #250 |
| 4 | GPT-5.1 (2025-11-13) | 79.09 | #105 |
| 5 | Gemma 4 31B (IT) | 78.79 | #178 |
| 6 | Llama 3.3 70B Instruct | 78.53 | #536 |
| 7 | GPT-5.4 (2026-03-05) | 78.44 | #58 |
| 8 | Llama 4 Maverick Instruct FP8 | 78.14 | #412 |
| 9 | Gemini 2.5 Flash (Non-reasoning) | 78.03 | #242 (Gemini 2.5 Flash) |
| 10 | Gemini 2.5 Flash Lite (Non-reasoning) | 77.63 | #421 (Gemini 2.5 Flash Lite) |
| 11 | Llama 4 Scout Instruct | 77.54 | #542 |
| 12 | Claude Haiku 4.5 (20251001) | 77.35 | #258 |
| 13 | Qwen 2.5 72B Instruct | 77.18 | #458 |
| 14 | Mistral Large 3 | 76.89 | #404 |
| 15 | Qwen 3 Next 80B A3B Instruct | 75.93 | #370 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-alghafa · How It Works · Data refreshed daily, snapshot 2026-10-23.