IslamicLegalBench — leaderboard
Islamic law and jurisprudence benchmark evaluating legal knowledge, reasoning, and public aggregate model performance.
Metric: Weighted Accuracy (%). Source: huggingface.co. Status: saturation imminent. 12 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 76.93 |
| 2 | Claude Sonnet 4.5 | 75 |
| 3 | Gemini 2.5 Pro | 73.57 |
| 4 | Grok 4 | 72.11 |
| 5 | DeepSeek R1 | 63.71 |
| 6 | Qwen 3 235B A22B 2507 Instruct | 60.31 |
| 7 | Llama 4 Maverick Instruct FP8 | 59.78 |
| 8 | GPT-OSS-120B | 46.54 |
| 9 | Llama 3.1 8B Instruct | 43.64 |
Interactive version: theaggregate.ai/benchmark?slug=islamiclegalbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.