PM-Bench: leaderboard
Metric: Set-F1 (%; F1 of the prospective actions the agent selects against the actions actually due, accumulated over the 80 steps and 81 scored tasks of the released simulated week, lures included; single agent with no memory scaffold, heartbeat or subagents; one run per model). Source: arxiv.org. Saturation forecast: Around December 2026. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.3 Codex | 78.9 |
| 2 | Mistral-Large-3-675B-Instruct-2512 | 75.3 |
| 3 | GPT-5.4 | 72.5 |
| 4 | Llama 3.3 70B Instruct | 64.4 |
| 5 | Mistral Small 3.2 | 52.6 |
| 6 | Qwen 3 32B | 51.4 |
| 7 | Qwen 3 14B | 43.1 |
| 8 | Qwen 3 8B | 42 |
Interactive version: theaggregate.ai/benchmark?slug=pm-bench · How It Works · Data refreshed daily, snapshot 2026-09-29.