Arabic Broad Leaderboard - Writing (incl Dialects): leaderboard
Arabic writing tasks, including dialects: the 22-question Writing (incl Dialects) category of SILMA's Arabic Broad Benchmark, whose answers are scored 0-10 by rules or an LLM judge. Category average score (0-10), higher is better; each model's latest result file is used.
Metric: Average Score (0-10). Source: huggingface.co. Saturation forecast: Estimated already saturated. 121 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4o | 8.55 | #357 |
| 2 | DeepSeek V4.1 Flash | 8.14 | #37 |
| 3 | Claude Sonnet 4 (20250514) | 8.09 | #223 |
| 4 | GPT-4.1 Mini | 8 | #363 |
| 5 | GPT-5.2 | 8 | #115 |
| 6 | Claude Opus 4.1 (20250805) | 8 | #139 |
| 7 | Gemini 2.0 Flash | 7.96 | #351 |
| 8 | Gemini 2.5 Pro (Preview 05-06) | 7.96 | #156 |
| 9 | GPT-5 | 7.91 | #105 |
| 10 | GPT-OSS-120B | 7.91 | #346 |
| 11 | Claude Opus 4.7 | 7.91 | #51 |
| 12 | Claude Fable 5 | 7.91 | #20 |
| 13 | Claude Opus 4 (20250514) | 7.91 | #164 |
| 14 | Qwen3.8 2.4T A95B | 7.91 | #59 |
| 15 | GPT-5 Mini (2025-08-07) | 7.86 | #175 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=arabic-broad-leaderboard-writing-incl-dialects · How It Works · Data refreshed daily, snapshot 2026-11-05.