Psi-Bench: leaderboard
Metric: Average persuasion effect over the three scenarios (1-9; DeepSeek-v3.2 judge, which also plays the persona-profiled client; three rounds after the client's opening message, client profile hidden from the model). Source: arxiv.org. Saturation forecast: Around 2029. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.1 | 5.79 |
| 2 | Qwen 3 Next 80B A3B | 5.38 |
| 3 | DeepSeek V4 Pro | 5.22 |
| 4 | Gemini 3.1 Pro (Preview) | 5.11 |
| 5 | GPT-5 Mini | 4.92 |
| 6 | Gemini 3 Flash | 4.89 |
| 7 | Qwen 3 32B | 4.54 |
| 8 | DeepSeek V3.2 | 4.47 |
| 9 | Grok 4 Fast | 4.19 |
| 10 | Qwen 3 8B | 4.05 |
Interactive version: theaggregate.ai/benchmark?slug=psi-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.