IslamicLegalBench - High Complexity: leaderboard
IslamicLegalBench high-complexity slice measures weighted accuracy on harder Islamic legal reasoning questions.
Metric: Weighted Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 12 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Sonnet 4.5 | 75.5 | #142 |
| 2 | Gemini 2.5 Pro | 73.12 | #152 |
| 3 | GPT-5 | 72.99 | #98 |
| 4 | Grok 4 | 72.88 | #181 |
| 5 | Qwen 3 235B A22B 2507 Instruct | 67.65 | #300 |
| 6 | DeepSeek R1 | 66.51 | #251 |
| 7 | Llama 4 Maverick Instruct FP8 | 55.14 | #414 |
| 8 | Llama 3.1 8B Instruct | 52.99 | #1035 |
| 9 | GPT-OSS-120B | 50.7 | #341 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=islamiclegalbench-high-complexity · How It Works · Data refreshed daily, snapshot 2026-10-20.