Visual Grounding on JARVIS-VLA Benchmark 1.0 (test)
88AccuracyJARVIS-VLA-Qwen2-VL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| JARVIS-VLA-Qwen2-VLModel Size=7B2025.03 | 88 | 1 | |
| Qwen2-VL (Grounding)Model Size=7B2025.03 | 63.6 | 2 | |
| Molmo-d-0924Model Size=7B2025.03 | 24.8 | 3 | |
| Qwen2-VL (Vision)Model Size=7B2025.03 | 19.8 | 4 | |
| Qwen2-VLModel Size=7B2025.03 | 16.6 | 5 | |
| Qwen2-VL (Knowledge)Model Size=7B2025.03 | 16.6 | 5 |