HELM Arabic - Alghafa - Multiple Choice Rating Sentiment Task: leaderboard
HELM Arabic (release v2.2.0) AlGhafa Multiple Choice Rating Sentiment Task task: the sentiment of rated reviews, including neutral, from the native AlGhafa Arabic multiple-choice benchmark; 1,000 items, exact match. Percent correct, 0-100, higher is better.
Metric: Exact Match (%). Source: crfm.stanford.edu. Saturation forecast: Around 2031. 41 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.6 | 63.1 | #94 |
| 2 | GPT-4.1 (2025-04-14) | 63 | #257 |
| 3 | GPT-5.1 (2025-11-13) | 61 | #107 |
| 4 | Claude Opus 4.7 | 60.2 | #51 |
| 5 | Gemma 4 31B (IT) | 60.2 | #178 |
| 6 | Llama 4 Maverick Instruct FP8 | 59.6 | #424 |
| 7 | Llama 3.3 70B Instruct | 57.9 | #541 |
| 8 | jais-adapted-7B-chat | 56.9 | #1160 |
| 9 | Llama 4 Scout Instruct | 56.4 | #547 |
| 10 | AceGPT-v2-8B-Chat | 56.4 | #746 |
| 11 | jais-adapted-70B-chat | 56.3 | #639 |
| 12 | Qwen 3.5 397B A17B | 56.1 | #154 |
| 13 | Gemini 2.5 Flash (Non-reasoning) | 56 | #255 (Gemini 2.5 Flash) |
| 14 | AceGPT-v2-70B-Chat | 56 | #421 |
| 15 | Claude Haiku 4.5 (20251001) | 55.8 | #264 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-alghafa-multiple-choice-rating-sentiment-task · How It Works · Data refreshed daily, snapshot 2026-11-05.