Physical AI Bench Understanding
Physical AI Bench understanding leaderboard for embodied physical reasoning over common sense, space, time, physics, robotics, autonomous driving, and video QA datasets.
25rows
overallprimary metric
2026-05-06sampled
Metadata
Metrics
Overall, Common Sense, Embodied Reasoning, Space, Time, Physics, BridgeData V2, RoboVQA, RoboFail, Agibot, HoloAssist, AV
| Rank | Subject | Overall | Model Match | Provenance | Sampled |
|---|---|---|---|---|---|
| 1 | Cosmos-Reason2-32B | 70.80 | — | Imported | 2026-05-06 |
| 2 | GPT-5 | 69.80 |  GPT-5 openai-gpt-5 | Imported | 2026-05-06 |
| 3 | Cosmos-Reason2-8B | 65.70 | — | Imported | 2026-05-06 |
| 4 | Qwen3-VL-235B-A22B-Instruct | 64.70 |  Qwen3 VL 235B A22B Instruct qwen-qwen3-vl-235b-a22b-instruct | Imported | 2026-05-06 |
| 5 | Qwen3-VL-235B-A22B-Thinking | 63.70 | Qwen3 VL 235B A22B Thinking qwen-qwen3-vl-235b-a22b-thinking | Imported | 2026-05-06 |
| 6 | Qwen3-VL-32B-Instruct | 62 | Qwen3 VL 32B Instruct qwen-qwen3-vl-32b-instruct | Imported | 2026-05-06 |
| 7 | Qwen3-VL-32B-Thinking | 61 | — | Imported | 2026-05-06 |
| 8 | Qwen2.5-VL-72B-Instruct | 60.80 | Qwen2.5 VL 72B Instruct qwen-qwen2.5-vl-72b-instruct | Imported | 2026-05-06 |
| 9 | Qwen3-VL-30B-A3B-Instruct | 59.50 | Qwen3 VL 30B A3B Instruct qwen-qwen3-vl-30b-a3b-instruct | Imported | 2026-05-06 |
| 10 | GLM-4.5V | 59.20 | GLM 4.5V z-ai-glm-4.5v | Imported | 2026-05-06 |
| 11 | Qwen3-VL-30B-A3B-Thinking | 57.30 | Qwen3 VL 30B A3B Thinking qwen-qwen3-vl-30b-a3b-thinking | Imported | 2026-05-06 |
| 12 | Qwen3-VL-8B-Thinking | 57.30 | Qwen3 VL 8B Thinking qwen-qwen3-vl-8b-thinking | Imported | 2026-05-06 |
| 13 | Qwen3-VL-8B-Instruct | 56.80 | Qwen3 VL 8B Instruct qwen-qwen3-vl-8b-instruct | Imported | 2026-05-06 |
| 14 | Cosmos-Reason2-2B | 56.40 | — | Imported | 2026-05-06 |
| 15 | InternVL3.5-241B-A28B | 56.30 | — | Imported | 2026-05-06 |
| 16 | GPT-4o | 56.20 | GPT-4o openai-gpt-4o | Imported | 2026-05-06 |
| 17 | InternVL3.5-38B | 55.80 | — | Imported | 2026-05-06 |
| 18 | Cosmos-Reason1-7B | 55.70 | — | Imported | 2026-05-06 |
| 19 | Qwen2.5-VL-32B-Instruct | 55.30 | — | Imported | 2026-05-06 |
| 20 | Qwen2.5-VL-7B-Instruct | 51 | — | Imported | 2026-05-06 |
| 21 | InternVL3.5-8B | 49.70 | — | Imported | 2026-05-06 |
| 22 | InternVL3.5-30B-A3B | 49.40 | — | Imported | 2026-05-06 |
| 23 | InternVL3.5-14B | 48.80 | — | Imported | 2026-05-06 |
| 24 | Qwen3-VL-2B-Instruct | 48.70 | — | Imported | 2026-05-06 |
| 25 | Claude-3.5-Sonnet | 46 |  Claude 3.5 Sonnet anthropic-claude-3.5-sonnet | Imported | 2026-05-06 |
No matching rows.