PlanningBench: leaderboard
Metric: All-pass (%) (self-reported). Source: benchmarklist.com. 18 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | GPT-5.4 (xHigh) | 63.17 |
| 2 | GPT-5.4 (High) | 58.56 |
| 3 | GPT-5.4 (Medium) | 58.09 |
| 4 | Seed 2.0 Pro (High) | 44.33 |
| 5 | DeepSeek V3.2 (Thinking) | 37.13 |
| 6 | Hy3-preview | 36.17 |
| 7 | Qwen 3.5 Plus (Thinking) | 34.03 |
| 8 | DeepSeek R1 | 33.07 |
| 9 | Gemini 2.5 Pro | 31.7 |
| 10 | DeepSeek V3.2 Exp | 29.27 |
| 11 | Seed 1.8 | 22.4 |
| 12 | Qwen 3 32B | 0.27 |
| 13 | Qwen 3 8B | 0 |
| 14 | Qwen 3 14B | 0 |
Interactive version: theaggregate.ai/benchmark?slug=planningbench · How It Works · Data refreshed daily, snapshot 2026-09-05.