RPCBench - Error Group - Inconsistent Premise - Proactive Detection: leaderboard
RPCBench tests proactive premise critique in recommendation requests. This series measures proactive detection on the Inconsistent Premise subset (1,867 questions), using the same eleven-model, three-judge experiment as the headline paper. Published PDR fractions from the authors' per-model CSV are scaled once to Detection Rate (%). Higher is better; domains and nested premise-error groups are related slices of the same experiment.
Metric: Detection Rate (%). Source: github.com. Saturation forecast: Around February 2027. 11 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V4 Pro | 75.95 | #113 |
| 2 | Claude Sonnet 4.6 | 74.99 | #97 |
| 3 | Qwen 3.5 397B A17B | 74.02 | #155 |
| 4 | Qwen 3.5 122B A10B | 73.75 | #186 |
| 5 | Gemini 3.1 Pro (Preview) | 73.59 | #66 |
| 6 | Qwen 3.5 Plus | 73.33 | #136 |
| 7 | DeepSeek V4 Flash | 69.31 | #156 |
| 8 | GPT-5.5 | 69.09 | #34 |
| 9 | Qwen 3.5 35B A3B | 68.61 | #270 |
| 10 | Llama 3.1 70B Instruct | 21.96 | #570 |
| 11 | Llama 3.1 8B Instruct | 17.46 | #1021 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-error-group-inconsistent-premise-proactive-detection · How It Works · Data refreshed daily, snapshot 2026-11-07.