MAWARITH: leaderboard
Metric: MIR-E score (0-1, times 100): 0.3 x heir identification + 0.3 x share assignment + 0.1 x adjustment (awl/radd, credited only when heirs and shares are fully correct) + 0.3 x final distribution, on the 500 held-out test cases of Arabic Islamic inheritance problems (calculator-generated, expert-annotated), zero-shot Arabic prompt asking for the full structured solution, temperature 0.1, 4,096-token context; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Flash | 90.1 | #237 |
| 2 | Qwen 3 32B (Thinking) | 43.7 | #424 (Qwen 3 32B) |
| 3 | GPT-OSS-120B | 39.1 | #330 |
| 4 | Fanar-C-2-27B | 32.8 | #614 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=mawarith · How It Works · Data refreshed daily, snapshot 2026-10-11.