HELM Arabic Enterprise - Arabic Finance Bool: leaderboard
The "Arabic Finance Bool" subtask of HELM Arabic Enterprise. Stanford HELM's Arabic Enterprise project: business-facing Arabic tasks — content generation, summarisation, translation, dialect handling and customer-support style requests — scored by rubric and judge metrics rather than exact match.
Metric: EM. Source: crfm.stanford.edu. Saturation forecast: Around July 2027. 35 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemma 4 31B (IT) | 82.46 | #178 |
| 2 | Claude Opus 4.7 | 78.95 | #50 |
| 3 | GPT-5.4 (2026-03-05) | 75.44 | #58 |
| 4 | GPT-5.1 (2025-11-13) | 71.93 | #105 |
| 5 | Claude Sonnet 4.6 | 70.18 | #92 |
| 6 | DeepSeek V4 Pro | 70.18 | #106 |
| 7 | Llama 3.3 70B Instruct | 68.42 | #536 |
| 8 | GPT-5.4 Mini (2026-03-17) | 66.67 | #171 |
| 9 | Qwen 3.5 397B A17B | 63.16 | #147 |
| 10 | Mistral Large 3 | 61.4 | #404 |
| 11 | Mistral Large 2 (Nov) Instruct (2411) | 61.4 | #468 |
| 12 | Gemini 2.5 Flash (Non-reasoning) | 59.65 | #242 (Gemini 2.5 Flash) |
| 13 | Gemini 2.5 Flash Lite (Non-reasoning) | 59.65 | #421 (Gemini 2.5 Flash Lite) |
| 14 | ALLaM-7B-Instruct-preview | 59.65 | #724 |
| 15 | Llama 4 Scout Instruct | 57.89 | #542 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=helm-arabic-enterprise-arabic-finance-bool · How It Works · Data refreshed daily, snapshot 2026-10-23.