RP-Bench - Flaw Hunter: leaderboard
RP-Bench Flaw Hunter measures whether models can identify defects in role-playing conversations and character behavior.
Metric: Flaw Hunter Score. Source: huggingface.co. Saturation forecast: Around January 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 72.1 | #63 |
| 2 | DeepSeek V3.2 | 68.8 | #205 |
| 3 | Claude Sonnet 4.5 | 67.1 | #142 |
| 4 | GLM-4.7 | 65.6 | #192 |
| 5 | GPT-4.1 | 65.5 | #248 |
| 6 | Gemini 2.5 Flash | 60 | #245 |
| 7 | Qwen 3.5 Flash | 58.3 | #232 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=rp-bench-flaw-hunter · How It Works · Data refreshed daily, snapshot 2026-10-20.