RHELM (RAG Top-20) - External Source: leaderboard
Metric: Accuracy (%; 249 external-source (attachment and email) questions, LLM judge, dialogue history plus external sources, dense top-20 retrieval). Source: arxiv.org. Saturation forecast: Around 2029. 4 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.5 | 33.7 |
| 2 | GPT-4.1 | 32.5 |
| 3 | Gemini 2.5 Pro | 31.3 |
| 4 | GPT-4.1 Mini | 28.5 |
Interactive version: theaggregate.ai/benchmark?slug=rhelm-rag-top-20-external-source · How It Works · Data refreshed daily, snapshot 2026-09-25.