DeepPlanning: leaderboard
Planning benchmark for complex multi-step tasks that require decomposing goals, tracking constraints, and revising a plan over a long reasoning horizon.
Metric: Score (%). Source: arxiv.org. Source: llm-stats.com. 10 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.7 Plus | 62.3 | #92 |
| 2 | Qwen 3.6 Plus | 41.5 | #120 |
| 3 | Qwen 3.5 397B A17B | 34.3 | #146 |
| 4 | Qwen 3 Max (Thinking) | 28.7 | #208 (Qwen 3 Max) |
| 5 | Qwen 3.6 35B A3B | 25.9 | #226 |
| 6 | Qwen 3.5 122B A10B | 24.1 | #173 |
| 7 | Qwen 3.5 35B A3B | 22.8 | #256 |
| 8 | Qwen 3.5 27B | 22.6 | #187 |
| 9 | Qwen 3.5 9B | 18 | #379 |
| 10 | Qwen 3.5 4B | 17.6 | #475 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=deepplanning · How It Works · Data refreshed daily, snapshot 2026-10-20.