BehaviorBench — leaderboard
Metric: Game Behav. Sim. (W â) (self-reported). Source: benchmarklist.com. 20 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 | 31.4 |
| 2 | Gemini 3.1 Flash | 31.3 |
| 3 | GPT-5.4 Mini | 29 |
| 4 | GPT-4.1 | 27.5 |
| 5 | Claude Opus 4.6 | 26 |
| 6 | Claude Sonnet 4.6 | 24.6 |
| 7 | Claude Haiku 4.5 | 24.5 |
| 8 | Llama 3.3 70B Instruct | 24.2 |
| 9 | Gemini 3.1 Pro (Preview) | 21.3 |
| 10 | DeepSeek V3.2 | 16.9 |
Interactive version: theaggregate.ai/benchmark?slug=behaviorbench · How the rankings work · Data refreshed daily, snapshot 2026-07-22.