Psi-Bench - Psychological Consultation: leaderboard
Psychological consultation in Psi-Bench (2026): the treatment effect of the model's counselling on the mindset of a simulated, persona-profiled client for CounselBench psychotherapy queries, rated 1 to 9 by a DeepSeek-V3.2 judge that also plays the client. Higher is better.
Metric: Treatment effect on the client's mindset, CounselBench psychotherapy queries (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 December 2027. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3 Next 80B A3B | 6.1 | #757 |
| 2 | DeepSeek V4 Pro | 6.04 | #188 |
| 3 | Gemini 3 Flash | 6.01 | #164 |
| 4 | Gemini 3.1 Pro (Preview) | 5.77 | #94 |
| 5 | GPT-5 Mini | 5.37 | #348 |
| 6 | GPT-5.1 | 5.37 | #229 |
| 7 | DeepSeek V3.2 | 5.25 | #406 |
| 8 | Qwen 3 32B | 5.1 | #960 |
| 9 | Grok 4 Fast | 4.94 | #441 |
| 10 | Qwen 3 8B | 4.53 | #1830 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=psi-bench-psychological-consultation · How It Works · Data refreshed daily, snapshot 2026-11-13.