SAHM - Event-Cause QA: leaderboard
Metric: Judge score (0-10) on the 80 event-cause reasoning questions over Arabic financial reports of SAHM, an Arabic financial and Sharia-compliance benchmark, each answer graded against the gold answer by a Gemini 2.5 Flash judge with an additive rubric, greedy decoding, mean of three runs; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Flash (Preview) | 9.84 |
| 2 | GPT-5 | 9.68 |
| 3 | Claude Opus 4.5 | 9.68 |
| 4 | Claude Sonnet 4.5 | 9.34 |
| 5 | Claude Haiku 4.5 | 9.11 |
| 6 | Gemma 3 27B (IT) | 8.72 |
| 7 | GPT-4o | 8.31 |
| 8 | Qwen 2.5 72B Instruct | 8.1 |
| 9 | GPT-4o Mini | 7.96 |
| 10 | Fanar-1-9B-Instruct | 7.59 |
| 11 | Qwen 2.5 14B Instruct | 7.5 |
| 12 | Gemma 3 4B (IT) | 7.41 |
| 13 | Gemma 2 9B (IT) | 7.14 |
| 14 | ALLaM-7B-Instruct-preview | 6.89 |
| 15 | Llama 3.1 70B Instruct | 6.62 |
Interactive version: theaggregate.ai/benchmark?slug=sahm-event-cause-qa · How It Works · Data refreshed daily, snapshot 2026-10-07.