PersTurnBench - Dissatisfaction Rate: leaderboard
PersTurnBench (2026) replays turns of real users conversations: the candidate model writes the next assistant reply and a verified personalized evaluator with user memory predicts how satisfied that user would be, calibrated per user. This board is the share in percent of replayed turns predicted to leave the user neutral or dissatisfied; lower is better.
Metric: Dissatisfied-turn rate (%): share of replay turns with a calibrated satisfaction of 3 or lower, PersTurnBench replay of 300 turns from 90 users (180 user-task blocks): the candidate fills only the current assistant turn and the frozen Qwen3-8B memory evaluator predicts the user's turn satisfaction, reference-CDF calibrated per user; lower is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 10 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.6 | 1.3 | #192 |
| 2 | DeepSeek V4 Pro | 2.7 | #196 |
| 3 | GLM-5.1 | 2.7 | #198 |
| 4 | Claude Opus 4.7 | 4.7 | #85 |
| 5 | GPT-5.5 | 5.7 | #54 |
| 6 | Gemini 3.1 Pro (Preview) | 6 | #99 |
| 7 | Qwen 3 8B | 8 | #1873 |
| 8 | DeepSeek V4 Flash | 10.7 | #267 |
| 9 | Gemini 3.1 Flash Lite | 13 | #371 |
| 10 | MiniMax-M2.7 | 13.7 | #539 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=persturnbench-dissatisfaction-rate · How It Works · Data refreshed daily, snapshot 2026-11-21.