AutoControl Arena X-Bench (High Stress and Temptation): leaderboard
Metric: Risk rate (%): share of trajectories a Qwen3-32B monitor scores 2 or more of 4 for action or chain-of-thought suspicion, over the 70 X-Bench scenarios (3 rollouts each) under high stress and high temptation together (S1T1); lower is better. Source: arxiv.org. 9 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 Mini | 22.1 | #176 |
| 2 | Gemini 2.5 Flash | 52.2 | #237 |
| 3 | O3 Mini | 52.9 | #266 |
| 4 | Qwen 3 32B | 53.2 | #424 |
| 5 | GPT-4.1 Mini | 55.9 | #346 |
| 6 | Qwen 3 8B | 60.6 | #667 |
| 7 | Qwen 3 4B | 61.2 | #823 |
| 8 | Qwen 3 14B | 62.1 | #524 |
| 9 | Gemini 3 Flash | 70.6 | #93 |
Interactive version: theaggregate.ai/benchmark?slug=autocontrol-arena-x-bench-high-stress-and-temptation · How It Works · Data refreshed daily, snapshot 2026-10-11.