RPCBench - Error Group - Unverifiable Premise - Proactive Detection: leaderboard

Metric: Detection Rate (%). Source: github.com. Saturation forecast: Around December 2026. 11 models tracked.

Top models

#ModelScore
1GPT-5.592.55
2DeepSeek V4 Pro91.08
3Gemini 3.1 Pro (Preview)89.61
4Qwen 3.5 122B A10B87.35
5DeepSeek V4 Flash84.71
6Qwen 3.5 35B A3B83.82
7Qwen 3.5 397B A17B82.55
8Qwen 3.5 Plus82.16
9Claude Sonnet 4.674.22
10Llama 3.1 70B Instruct41.57
11Llama 3.1 8B Instruct35.69

Interactive version: theaggregate.ai/benchmark?slug=rpcbench-error-group-unverifiable-premise-proactive-detection · How It Works · Data refreshed daily, snapshot 2026-09-24.