HalluScore - Arabic Culture: leaderboard
HalluScore (2026) is an Arabic question-answering benchmark of 827 hallucination-prone Modern Standard Arabic questions, with human annotation of model answers. This board is the factual hallucination rate, the percent of answers that state an unsupported fact with confidence on the 308 questions about Arab culture, zero-shot at temperature 0 (expressed uncertainty or refusal does not count); lower is better.
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: Around April 2027. 17 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | GPT-5 | 33.44 | #178 | |
| 2 | Claude Opus 4 | 45.45 | #326 | |
| 3 | GPT-4o | 47.08 | #751 | |
| 4 | Claude Sonnet 4.5 | 52.27 | #261 | |
| 5 | O4 Mini | 59.74 | #368 | |
| 6 | Llama 4 Maverick Instruct FP8 | 63.31 | #1023 | |
| 7 | DeepSeek V3 | 66.23 | #729 | |
| 8 | ALLaM-7B-Instruct-preview | 69.16 | #2001 | |
| 9 | Grok 4 Fast | Reasoning | 69.16 | #447 |
| 10 | DeepSeek R1 | 69.81 | #482 | |
| 11 | HalluScore Jais-6.7B (checkpoint unspecified) | 75.97 | ||
| 12 | Qwen3-235B-A22B-Instruct-2507-tput | 76.95 | #548 | |
| 13 | Qwen 3 Next 80B A3B Instruct | 77.6 | #778 | |
| 14 | Grok 4 | 79.22 | #319 | |
| 15 | Fanar-1-9B | 84.09 | #1913 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=halluscore-arabic-culture · How It Works · Data refreshed daily, snapshot 2026-11-21.