PreAct-Bench: leaderboard
Metric: Prefix Foresight F1 with the full prefix (PFF(100), binary F1 of the unethical class, 0-1 scaled to %): given an agent trajectory truncated just before its first overt unethical step, the monitor predicts whether the trajectory will turn unethical; paired ethical and unethical trajectories per task; default generation parameters; higher is better; mean over the five domains (200 trajectories each). Source: arxiv.org. Saturation forecast: Around June 2028. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemma 3 27B | 83.9 |
| 2 | GPT-5.4 | 81.5 |
| 3 | Qwen 3 30B A3B 2507 Instruct | 79.6 |
| 4 | Qwen 3 8B (Non-reasoning) | 72.7 |
| 5 | Gemma 3 4B | 66.7 |
| 6 | Qwen 3 0.6B (Non-reasoning) | 66.7 |
| 7 | Qwen 3 8B (Thinking) | 62.7 |
| 8 | GPT-4o | 61.3 |
| 9 | Qwen 3 30B A3B 2507 (Thinking) | 58.8 |
| 10 | Qwen 3 0.6B (Thinking) | 58.5 |
| 11 | Qwen 3 4B 2507 Instruct | 51 |
| 12 | Qwen 3 4B 2507 (Thinking) | 49.3 |
| 13 | GPT-OSS-20B | 49.1 |
| 14 | GPT-OSS-120B | 45.5 |
Interactive version: theaggregate.ai/benchmark?slug=preact-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.