RPCBench - Error Group - Boundary-Violating Premise - Evidence Distortion: leaderboard
Metric: Evidence Distortion Rate (%). Source: github.com. Saturation forecast: Estimated already saturated. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Plus | 0.92 |
| 2 | Qwen 3.5 397B A17B | 1.61 |
| 3 | GPT-5.5 | 2.75 |
| 4 | Qwen 3.5 122B A10B | 3.21 |
| 5 | Gemini 3.1 Pro (Preview) | 4.13 |
| 6 | Qwen 3.5 35B A3B | 6.19 |
| 7 | Claude Sonnet 4.6 | 8.03 |
| 8 | DeepSeek V4 Flash | 19.5 |
| 9 | DeepSeek V4 Pro | 20.41 |
| 10 | Llama 3.1 70B Instruct | 22.02 |
| 11 | Llama 3.1 8B Instruct | 23.39 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-error-group-boundary-violating-premise-evidence-distortion · How It Works · Data refreshed daily, snapshot 2026-09-24.