HalluScore - Arabic Culture: leaderboard
Metric: Factual hallucination rate (%) on the 308 culturally grounded HalluScore questions about the Arab world, zero-shot short-answer Arabic question answering at temperature 0 with no retrieval, responses annotated by humans; an answer that states an unsupported fact with confidence counts, an expressed uncertainty or refusal does not; lower is better. Source: arxiv.org. Saturation forecast: Not forecast. 17 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5 | 33.44 |
| 2 | Claude Opus 4 | 45.45 |
| 3 | GPT-4o | 47.08 |
| 4 | Claude Sonnet 4.5 | 52.27 |
| 5 | O4 Mini | 59.74 |
| 6 | Llama 4 Maverick Instruct FP8 | 63.31 |
| 7 | DeepSeek V3 | 66.23 |
| 8 | ALLaM-7B-Instruct-preview | 69.16 |
| 9 | Grok 4 Fast (Reasoning) | 69.16 |
| 10 | DeepSeek R1 | 69.81 |
| 11 | Qwen 3 Next 80B A3B Instruct | 77.6 |
| 12 | Grok 4 | 79.22 |
| 13 | Fanar-1-9B | 84.09 |
| 14 | O3 | 86.36 |
Interactive version: theaggregate.ai/benchmark?slug=halluscore-arabic-culture · How It Works · Data refreshed daily, snapshot 2026-10-07.