RPCBench - Error Group - Inconsistent Premise - Evidence Faithfulness: leaderboard
Metric: Faithfulness Index (%). Source: github.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Plus | 93.47 |
| 2 | Qwen 3.5 397B A17B | 92.85 |
| 3 | Qwen 3.5 122B A10B | 92.66 |
| 4 | GPT-5.5 | 92.21 |
| 5 | Qwen 3.5 35B A3B | 90.6 |
| 6 | Gemini 3.1 Pro (Preview) | 89.29 |
| 7 | DeepSeek V4 Pro | 85.97 |
| 8 | DeepSeek V4 Flash | 84.92 |
| 9 | Claude Sonnet 4.6 | 84.49 |
| 10 | Llama 3.1 70B Instruct | 39.1 |
| 11 | Llama 3.1 8B Instruct | 32.16 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-error-group-inconsistent-premise-evidence-faithfulness · How It Works · Data refreshed daily, snapshot 2026-09-24.