RPCBench - Composite Premise Critique: leaderboard
Metric: Composite Score (0-100). Source: arxiv.org. Saturation forecast: Around January 2027. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.5 Plus | 52.61 |
| 2 | Qwen 3.5 397B A17B | 52.59 |
| 3 | Qwen 3.5 122B A10B | 52.45 |
| 4 | GPT-5.5 | 51.8 |
| 5 | DeepSeek V4 Pro | 51.66 |
| 6 | Claude Sonnet 4.6 | 50.92 |
| 7 | Gemini 3.1 Pro (Preview) | 50.47 |
| 8 | Qwen 3.5 35B A3B | 48.51 |
| 9 | DeepSeek V4 Flash | 46.55 |
| 10 | Llama 3.1 70B Instruct | 13.59 |
| 11 | Llama 3.1 8B Instruct | 10.19 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-composite-premise-critique · How It Works · Data refreshed daily, snapshot 2026-09-24.