BehaviorBench - Survey Response from Demographics (Distribution): leaderboard
Metric: Wasserstein distance between predicted and observed answer distributions per item (five-point scale, so 0 to 4), averaged over the 50 items, predicting from demographics on the Big Five personality survey responses (50 five-point items) of held-out subjects; lower is better. Source: arxiv.org. Saturation forecast: Around December 2027. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 0.75 |
| 2 | DeepSeek V3.2 | 0.76 |
| 3 | GPT-4.1 | 0.79 |
| 4 | Gemini 3.1 Flash Lite (Preview) | 0.82 |
| 5 | Claude Sonnet 4.6 | 0.86 |
| 6 | Claude Opus 4.6 | 0.86 |
| 7 | GPT-5.4 (High) | 0.9 |
| 8 | GPT-5.4 Mini (High) | 0.91 |
| 9 | Claude Haiku 4.5 | 0.93 |
| 10 | Llama 3.3 70B Instruct | 1.02 |
| 11 | Qwen 3 4B | 1.14 |
Interactive version: theaggregate.ai/benchmark?slug=behaviorbench-survey-response-from-demographics-distribution · How It Works · Data refreshed daily, snapshot 2026-09-29.