When Simulation Lies: leaderboard
Metric: Pert Acc (self-reported). Source: benchmarklist.com. 22 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | O4 Mini | 50.2 |
| 2 | DeepSeek R1 Distill Qwen 14B | 44.24 |
| 3 | Qwen 3.5 9B | 43.7 |
| 4 | Qwen 2.5 3B Instruct | 39.01 |
| 5 | Qwen 2.5 1.5B Instruct | 36.68 |
| 6 | Llama 3.2 3B Instruct | 34.1 |
Interactive version: theaggregate.ai/benchmark?slug=when-simulation-lies · How It Works · Data refreshed daily, snapshot 2026-09-05.