Kimi Hallu Bench
Kimi K3.1 Pro 8.30% In use
Models answer hard prompts drawn from practical work tasks. Fact-checking agents score the share of claims that are hallucinated.
In-house. Moonshot reports a set without tools and a Hard set with tools. Moonshot says the rates are a stress test and not daily error rates.
Made by Moonshot AI Released Oct 2026 Hallucination rates Tasks kept private Last in a launch post Oct 2026
Last model added Oct 2026
- Level
- Expert Specialists with years of training, often over hours of work
- Scored
- Graded by a model A language model scores each response, usually against a rubric.
- Answers
- Open-ended Free-form text, documents, designs or media judged as a whole.
1
Model scored in lab cards
1
Launch post lists it, from 1 lab
Kimi K3.1 Pro leads with 8.30%
Best score per model as labs report them in their own cards. Labs pick their own settings, so these compare less cleanly than independent runs.
- 1 Kimi K3.1 Pro 8.30%
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | Kimi K3.1 Pro | 8.30% | Oct 2026 | Model card |
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Reported by the lab in its model card
One lab put it in a launch post in the last six months
Launch posts from every lab that has listed it, one tick per post. Tall colored ticks include it, and short gray ones leave it out.
| Date | Launch post | Score | As the post writes it |
|---|---|---|---|
| 24 Oct 2026 | Kimi K3.1 Pro | 8.3 | Kimi Hallu Bench Hard w/ tools (Internal, lower is better) |
More in Factuality
- SimpleVQA In use Models answer 2,025 short factual questions about images across 9 tasks and 9 topics. An LLM judge grades answers against fixed references.
Qwen3.7-Plus 81.7% - HallusionBench In use Models answer 1,129 yes or no questions about 346 images, built in pairs to expose language hallucination and visual illusion. Scored by accuracy.
Keye-VL-2.0-30B-A3B 73.5% - SimpleQA Verified In use 1,000 human-verified, de-duplicated short-answer factual questions measuring parametric knowledge.
Gemini 3.1 Pro 77.5% - GaRAGe In use Models write long-form RAG answers to 2,366 questions with human-annotated grounding passages. Scoring checks whether answers use the relevant passages.
Nemotron 3.5 Bolt E4B 40.7% - RAGTruth In use A word-level annotated corpus of hallucinations in retrieval-augmented generation answers. It is used to train and evaluate hallucination detection and RAG faithfulness. Nemotron 3.5 Bolt E4B 77.9%
- TruthfulQA Fading Models answer 817 questions across 38 categories that are written to elicit common misconceptions. Scoring checks for truthful answers, often in multiple-choice form.
MAI-Thinking-1 88%