PLAN-Bench - Overall: leaderboard
PLAN-Bench - Overall. 699 embodied visual questions; source task partition, nearest tenth percent. Reported metric: Accuracy (%). Primary published table, with source configuration boundaries preserved.
Metric: Accuracy (%). Source: arxiv.org. Saturation forecast: Around January 2027. 28 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.8 Max | 75 | #31 |
| 2 | Kimi K3 | 70.2 | #30 |
| 3 | GPT-5.6 Sol | 65.8 | #18 |
| 4 | Qwen 3.8 27B | 62.2 | #96 |
| 5 | Qwen 3.5 27B | 55.9 | #183 |
| 6 | Claude Opus 4.8 | 51.4 | #38 |
| 7 | Qwen 3.5 35B A3B | 49.5 | #250 |
| 8 | Qwen 3.5 9B | 49.2 | #367 |
| 9 | Gemini 3.1 Pro (Preview) | 46.4 | #56 |
| 10 | Gemma 4 31B (IT) | 43.5 | #172 |
| 11 | Qwen 3.5 4B | 43.2 | #463 |
| 12 | Step3 VL 10B | 35.9 | #477 |
| 13 | Gemma 4 12B (IT) | 28.5 | #356 |
| 14 | Qwen 3.5 2B | 20.2 | #1109 |
| 15 | Qwen 3 VL 8B Instruct | 17.9 | #414 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #7 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=plan-bench-overall · How It Works · Data refreshed daily, snapshot 2026-10-16.