PersonaArena: leaderboard
Large language models (LLMs) increasingly serve as interactive social agents, yet their ability to maintain coherent and authentic persona-level role-playing remains limited, particularly in realistic social scenarios. Existing research predominantly focuses on character-level settings and relies on static evaluation formats, failing to capture the complexity of everyday social interactions. In th
Metric: Average (self-reported). Source: benchmarklist.com. Saturation forecast: Around January 2027. 18 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.1 | 3.96 | #135 |
| 2 | GPT-4.1 | 3.95 | #248 |
| 3 | DeepSeek V3.2 | 3.9 | #205 |
| 4 | Qwen 3 32B | 3.81 | #443 |
| 5 | Mistral Small 3.2 | 3.75 | #605 |
| 6 | GPT-4o | 3.74 | #343 |
| 7 | GPT-4o Mini | 3.72 | #599 |
| 8 | Grok 3 | 3.63 | #304 |
| 9 | Qwen 3 14B | 3.59 | #537 |
| 10 | Phi-4 | 3.49 | #706 |
| 11 | GPT-3.5 | 3.46 | #818 |
| 12 | Qwen 3 8B | 3.36 | #674 |
| 13 | Llama 3.1 8B | 3.33 | #1151 |
| 14 | Llama 3.2 3B | 3.29 | #1396 |
| 15 | Qwen 3 4B | 3.28 | #829 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=personaarena · How It Works · Data refreshed daily, snapshot 2026-10-20.