Kluster Hallucination Detection - RAG Method 3 Resistance: leaderboard
The "RAG Method 3 Resistance" subtask of Kluster Hallucination Detection. A task from the Kluster Hallucination Detection benchmark, measuring how reliably a model detects unsupported or fabricated claims.
Metric: Resistance (100 - Hallucination Rate %). Source: huggingface.co. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek R1 0528 | 96.42 | #222 |
| 2 | Llama 4 Maverick Instruct FP8 | 94.28 | #414 |
| 3 | Qwen 3 235B A22B FP8 | 93.37 | #301 |
| 4 | Llama 4 Scout Instruct | 93.02 | #539 |
| 5 | DeepSeek V3 (0324) | 92.91 | #340 |
| 6 | Mistral Small 3 | 92.91 | #627 |
| 7 | Magistral Small | 88.13 | #1016 |
| 8 | Qwen 2.5 VL 7B Instruct | 86.76 | #654 |
| 9 | Mistral Nemo Instruct (2407) | 85.58 | #936 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=kluster-hallucination-detection-rag-method-3-resistance · How It Works · Data refreshed daily, snapshot 2026-10-20.