RPCBench - Domain - MIND-small - Evidence Faithfulness: leaderboard
Metric: Faithfulness Index (%). Source: github.com. Saturation forecast: Around December 2026. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 122B A10B | 86.42 |
| 2 | GPT-5.5 | 86.06 |
| 3 | Qwen 3.5 397B A17B | 86.06 |
| 4 | Qwen 3.5 35B A3B | 85.78 |
| 5 | Qwen 3.5 Plus | 83.93 |
| 6 | Gemini 3.1 Pro (Preview) | 79.37 |
| 7 | DeepSeek V4 Pro | 78.73 |
| 8 | DeepSeek V4 Flash | 76.88 |
| 9 | Claude Sonnet 4.6 | 72.05 |
| 10 | Llama 3.1 70B Instruct | 42.89 |
| 11 | Llama 3.1 8B Instruct | 33.14 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-domain-mind-small-evidence-faithfulness · How It Works · Data refreshed daily, snapshot 2026-09-24.