SAHM - Event-Cause QA: leaderboard

Open-ended Arabic financial reasoning on SAHM (2026): the 80 event-cause reasoning questions over Arabic financial reports, each answer scored from 0 to 10 by a Gemini 2.5 Flash judge against an expert-verified reference with an additive rubric. Higher is better.

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. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Gemini 3 Flash (Preview)9.84#152
2GPT-59.68#178
3Claude Opus 4.59.68#160
4Claude Sonnet 4.59.34#261
5Claude Haiku 4.59.11#583
6Gemma 3 27B (IT)8.72#1324
7GPT-4o8.31#751
8Qwen 2.5 72B Instruct8.1#1117
9GPT-4o Mini7.96#1667
10Fanar-1-9B-Instruct7.59#1844
11Qwen 2.5 14B Instruct7.5#1775
12Gemma 3 4B (IT)7.41#4139
13Gemma 2 9B (IT)7.14#2369
14ALLaM-7B-Instruct-preview6.89#2001
15Llama 3.1 70B Instruct6.62#1571

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=sahm-event-cause-qa · How It Works · Data refreshed daily, snapshot 2026-11-21.