Action-conditioned reasoning on MindCube
36.85Accuracy
World2VLM-GRPO
Updated 1mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.04 | 36.85 | |
2026.04 | 36.57 | |
2026.04 | 34.86 | |
2026.04 | 33.85 | |
2026.04 | 33.14 | |
2026.04 | 29.26 |
World2VLM-GRPO
| Method | Links | |
|---|---|---|
2026.04 | 36.85 | |
2026.04 | 36.57 | |
2026.04 | 34.86 | |
2026.04 | 33.85 | |
2026.04 | 33.14 | |
2026.04 | 29.26 |