RAGTruth: leaderboard
Retrieval-augmented generation hallucination corpus for evaluating grounded factuality and hallucination detection in RAG outputs.
Metric: RAG Truth (imported). Source: github.com. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude 3.5 Sonnet | 86.1 | #350 |
| 2 | Mistral Large 2 (Jul) | 85.9 | #484 |
| 3 | GPT-4o (2024-05-13) | 84.3 | #325 |
| 4 | Llama 3.1 405B Instruct | 82.9 | #457 |
| 5 | Llama 3.3 70B Instruct | 82.6 | #534 |
| 6 | Qwen 2.5 72B Instruct | 81.9 | #454 |
| 7 | QwQ 32B-Preview | 72.4 | #651 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=ragtruth · How It Works · Data refreshed daily, snapshot 2026-10-20.