When Simulation Lies — leaderboard
Metric: Pert Acc (self-reported). Source: benchmarklist.com. 24 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 14B | 52.9 |
| 2 | Qwen 3 32B | 52.4 |
| 3 | O4 Mini | 50.2 |
| 4 | Qwen 3 8B | 44.8 |
| 5 | DeepSeek R1 Distill Qwen 14B | 44.2 |
| 6 | Qwen 3.5 9B | 43.7 |
| 7 | Qwen 2.5 3B Instruct | 39 |
| 8 | Qwen 2.5 1.5B Instruct | 36.7 |
| 9 | Llama 3.2 3B Instruct | 34.1 |
Interactive version: theaggregate.ai/benchmark?slug=when-simulation-lies · How the rankings work · Data refreshed daily, snapshot 2026-07-22.