Embodied Manipulation on EB-Manipulation
42Average ScoreClaude-3.7
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Claude-3.7Tool use=Yes2026.05 | 42 | 46 | 44 | 50 | 36 | 33 | |
| Claude-3.7Tool use=No2026.05 | 38 | 33 | 38 | 46 | 33 | 38 | |
| Claude-4.6Tool use=Yes2026.05 | 38 | 40 | 35 | 42 | 40 | 31 | |
| GPT-4oTool use=Yes2026.05 | 37 | 44 | 31 | 40 | 36 | 35 | |
| Claude-4.6Tool use=No2026.05 | 37 | 40 | 35 | 40 | 36 | 37 | |
| GPT-5Tool use=Yes2026.05 | 35 | 42 | 27 | 44 | 31 | 31 | |
| GPT-4oTool use=No2026.05 | 31 | 31 | 29 | 27 | 36 | 33 | |
| Qwen2.5-32BTool use=Yes2026.05 | 29 | 27 | 35 | 33 | 21 | 28 | |
| GPT-5Tool use=No2026.05 | 27 | 21 | 25 | 31 | 22 | 35 | |
| Qwen3-8BTool use=Yes2026.05 | 27 | 29 | 35 | 31 | 19 | 19 | |
| Qwen3.5-35BTool use=Yes2026.05 | 27 | 33 | 29 | 27 | 22 | 23 | |
| Qwen3-32BTool use=Yes2026.05 | 25 | 27 | 25 | 31 | 17 | 22 | |
| Qwen3-32BTool use=No2026.05 | 20 | 21 | 25 | 17 | 19 | 19 | |
| Qwen3-8BTool use=No2026.05 | 19 | 15 | 15 | 25 | 6 | 31 | |
| Qwen2.5-32BTool use=No2026.05 | 17 | 13 | 13 | 23 | 11 | 25 | |
| Qwen3.5-35BTool use=No2026.05 | 15 | 13 | 17 | 15 | 17 | 17 |