RP-Bench - Elo: leaderboard
RP-Bench Elo rating measures role-playing model performance from pairwise preference comparisons across character consistency and interaction quality.
Metric: Elo. Source: huggingface.co. Saturation forecast: Around May 2027. 8 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Claude Opus 4.6 | 1705.7 | #63 |
| 2 | DeepSeek V3.2 | 1638.3 | #205 |
| 3 | Claude Sonnet 4.5 | 1541.1 | #142 |
| 4 | GPT-4.1 | 1522.7 | #248 |
| 5 | GLM-4.7 | 1491.9 | #192 |
| 6 | Gemini 2.5 Flash | 1407.8 | #245 |
| 7 | Qwen 3.5 Flash | 1332.5 | #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-elo · How It Works · Data refreshed daily, snapshot 2026-10-20.