Best LLM for
Factual Accuracy
Vectara HHEM measures summarization faithfulness; TruthfulQA tests resistance to common misconceptions. SimpleQA adds vendor-published factual-recall evidence. A factuality winner is named when 20 callable models have current independent evidence.
No current winner is published: qualifying independent evidence is older than 30 days. The table below shows available corroborated evidence, not a publishable current winner.
| # | Model | Factuality | Reasoning | Price/M |
|---|---|---|---|---|
| 1 | Meta: Llama 3.3 70B Instruct | 94.9 | 60.4 | $0.10/M |
| 2 | OpenAI: GPT-5.4 Mini | 93.5 | — | $0.75/M |
| 3 | Qwen: Qwen3 32B | 93.1 | — | $0.08/M |
| 4 | DeepSeek: DeepSeek V3.2 | 92.7 | 54.6 | $0.23/M |
| 5 | OpenAI: GPT-4.1 | 92.6 | 63.1 | $2.00/M |
| 6 | Google: Gemini 2.5 Pro | 92.0 | 53.2 | $1.25/M |
| 7 | OpenAI: GPT-5.4 | 92.0 | 72.9 | $2.50/M |
| 8 | Google: Gemma 3 27B | 91.6 | 60.0 | $0.08/M |
| 9 | Google: Gemini 2.5 Flash | 91.2 | 29.2 | $0.30/M |
| 10 | DeepSeek: DeepSeek V4 Pro | 90.4 | — | $0.43/M |
| 11 | OpenAI: GPT-5.5 | 89.7 | 74.3 | $5.00/M |
| 12 | Qwen: Qwen3 235B A22B | 89.7 | — | $0.45/M |
| 13 | Anthropic: Claude Sonnet 4 | 88.7 | 67.4 | $3.00/M |
| 14 | DeepSeek: R1 | 87.7 | 36.4 | $0.70/M |
| 15 | Qwen2.5 Coder 32B Instruct | 53.2 | — | $0.66/M |
A 97% versus 94% factual consistency rate implies 3% versus 6% inconsistent summaries on that measure: twice the failure rate, not a cosmetic difference.
If deployment constraints require an open-weight model, compare its current measured factuality row against the leader. Do not infer factuality from weight availability or vendor claims.
HHEM tests summarization faithfulness; SimpleQA tests factual recall. A model that scores well on one may not score well on the other. Always evaluate on your specific task — a model that rarely hallucinates in summaries may still confabulate on open-domain questions.
Low hallucination matters most for legal document review, medical information, RAG pipelines, and any application where invented facts cause real harm. For creative writing or brainstorming, hallucination score is irrelevant — pick by writing or EQ score instead.
Which LLM hallucinates the least?
No broad factuality winner is published yet. Vectara HHEM is independent evidence; SimpleQA is vendor-published and cannot provide the current two-source corroboration required across 20 callable models.
What is Vectara HHEM?
HHEM (Hughes Hallucination Evaluation Model) is an independent benchmark by Vectara that measures how often LLMs introduce factual inconsistencies when summarizing documents. Models are scored on Factual Consistency Rate (0-100%), where higher means fewer hallucinations.
How reliable is hallucination benchmarking?
Hallucination benchmarks measure different failure modes. Vectara HHEM focuses on summarization faithfulness — whether a model invents details not in the source document. SimpleQA measures factual recall on closed-domain questions. Neither captures all hallucination types. Models that score well on summarization may still hallucinate on open-domain questions or multi-step reasoning tasks.