RHELM (RAG Top-20) - Hallucination: leaderboard
RHELM category (Zhang et al., Microsoft and Renmin University of China, 2026) that poses questions built on misattributed or fabricated details, or on requests that contradict the persona's stated preferences or current state, which the assistant must detect and correct. The model answers from the top 20 retrieved chunks of dialogue history and external documents; accuracy in percent on 197 questions judged 0 or 1 by an LLM, higher is better.
Metric: Accuracy (%; 197 hallucination questions, LLM judge, dialogue history plus external sources, dense top-20 retrieval). Source: arxiv.org. Saturation forecast: Around February 2027. 4 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 64.5 | #262 |
| 2 | Claude Opus 4.5 | 60.9 | #152 |
| 3 | GPT-4.1 | 19.8 | #492 |
| 4 | GPT-4.1 Mini | 13.2 | #775 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=rhelm-rag-top-20-hallucination · How It Works · Data refreshed daily, snapshot 2026-11-12.