BehaviorBench - Sequential Survey Response: leaderboard

Metric: Accuracy (%) of predicting the next five-point answer of a subject from their earlier answers in survey order, on the Big Five personality survey responses (50 five-point items) of held-out subjects; chance 20; higher is better. Source: arxiv.org. Saturation forecast: Around 2034. 18 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)42.5
2GPT-5.4 (High)42.4
3Claude Sonnet 4.641.8
4Claude Opus 4.641.8
5GPT-5.4 Mini (High)41.7
6Gemini 3.1 Flash Lite (Preview)40.5
7Claude Haiku 4.540.2
8GPT-4.139
9Llama 3.3 70B Instruct37.7
10Qwen 3 4B37
11DeepSeek V3.236.2

Interactive version: theaggregate.ai/benchmark?slug=behaviorbench-sequential-survey-response · How It Works · Data refreshed daily, snapshot 2026-09-29.