RPCBench - Error Group - Underspecified Premise - Composite Premise Critique: leaderboard
Metric: Composite Score (0-100). Source: github.com. Saturation forecast: Around 2032. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | DeepSeek V4 Pro | 9.13 |
| 2 | Claude Sonnet 4.6 | 7.32 |
| 3 | Qwen 3.5 35B A3B | 7.32 |
| 4 | Gemini 3.1 Pro (Preview) | 7.19 |
| 5 | Qwen 3.5 122B A10B | 6.57 |
| 6 | DeepSeek V4 Flash | 6.45 |
| 7 | Qwen 3.5 Plus | 6.32 |
| 8 | Qwen 3.5 397B A17B | 6.06 |
| 9 | GPT-5.5 | 4.79 |
| 10 | Llama 3.1 8B Instruct | 2.55 |
| 11 | Llama 3.1 70B Instruct | 1.77 |
Interactive version: theaggregate.ai/benchmark?slug=rpcbench-error-group-underspecified-premise-composite-premise-critique · How It Works · Data refreshed daily, snapshot 2026-09-24.