Psi-Bench - Viewpoint Debate: leaderboard
Metric: Persuasion effect on the client's opinion, Change My View debates (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 May 2027. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.1 | 6.12 |
| 2 | Gemini 3.1 Pro (Preview) | 5.89 |
| 3 | DeepSeek V4 Pro | 5.71 |
| 4 | GPT-5 Mini | 5.27 |
| 5 | DeepSeek V3.2 | 4.81 |
| 6 | Gemini 3 Flash | 4.68 |
| 7 | Grok 4 Fast | 4.57 |
| 8 | Qwen 3 32B | 4.3 |
| 9 | Qwen 3 Next 80B A3B | 4.2 |
| 10 | Qwen 3 8B | 3.51 |
Interactive version: theaggregate.ai/benchmark?slug=psi-bench-viewpoint-debate · How It Works · Data refreshed daily, snapshot 2026-09-29.