Maze Navigation on Maze Navigation (val)
4.45Human ScoresketchVLM
Evaluation results
| Method | Links | |
|---|---|---|
| sketchVLMvariant=Standard (Square icon)2026.04 | 4.45 | |
| sketchVLMvariant=Star icon variant2026.04 | 4.36 | |
| Baseline Model (Multi-colored star)icon=multi-colored star2026.04 | 4.29 | |
| Baseline Model (Blue circle)icon=blue circle2026.04 | 1.96 | |
| Baseline Model (OpenAI icon)icon=multi-color circle logo2026.04 | 1.25 |