HalluScore - Factual Hallucination: 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, zero-shot at temperature 0 (expressed uncertainty or refusal does not count); lower is better.

Metric: Factual hallucination rate (%): share of the 827 HalluScore questions (hallucination-prone Modern Standard Arabic questions across eleven types and many domains), 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 February 2027. 17 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-525.15#178
2Claude Opus 433.01#326
3Claude Sonnet 4.535.79#261
4GPT-4o41.23#751
5O4 Mini46.43#368
6Grok 4 FastReasoning52.96#447
7DeepSeek V354.78#729
8Llama 4 Maverick Instruct FP857.56#1023
9DeepSeek R157.68#482
10Qwen3-235B-A22B-Instruct-2507-tput59.49#548
11Qwen 3 Next 80B A3B Instruct61.43#778
12Grok 465.42#319
13ALLaM-7B-Instruct-preview68.68#2001
14Fanar-1-9B79.32#1913
15O380.05#237

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

Interactive version: theaggregate.ai/benchmark?slug=halluscore-factual-hallucination · How It Works · Data refreshed daily, snapshot 2026-11-21.