Vision-Language Understanding on Winoground (test)
76.75Text ScoreLLaVA-OV
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| LLaVA-OVModel Family=LLaVA-OV, Model Size=≥7B (7B), Evaluation State=Frozen2025.08 | 76.75 | 36.75 | 33 | |
| LLaVA-OV + FOCUSModel Family=LLaVA-OV, Model Size=≥7B (7B), Evaluation State=Frozen2025.08 | 76.75 | 48.5 | 42.25 | |
| Qwen2.5-VLModel Family=Qwen2.5-VL, Model Size=≥7B (7B), Evaluation State=Frozen2025.08 | 74.5 | 39.75 | 34 | |
| Qwen2.5-VL + FOCUSModel Family=Qwen2.5-VL, Model Size=≥7B (7B), Evaluation State=Frozen2025.08 | 74.5 | 58.5 | 50.75 | |
| InternVL3Model Family=InternVL3, Model Size=≥7B (8B), Evaluation State=Frozen2025.08 | 70.75 | 40.75 | 34.25 | |
| InternVL3 + FOCUSModel Family=InternVL3, Model Size=≥7B (8B), Evaluation State=Frozen2025.08 | 69.25 | 42.25 | 35.75 | |
| Qwen2.5-VLModel Family=Qwen2.5-VL, Model Size=<7B (3B), Evaluation State=Frozen2025.08 | 56.25 | 36.75 | 26 | |
| Qwen2.5-VL + FOCUSModel Family=Qwen2.5-VL, Model Size=<7B (3B), Evaluation State=Frozen2025.08 | 56.25 | 36.5 | 26 | |
| InternVL3Model Family=InternVL3, Model Size=<7B (2B), Evaluation State=Frozen2025.08 | 47.25 | 6.25 | 4.5 | |
| InternVL3 + FOCUSModel Family=InternVL3, Model Size=<7B (2B), Evaluation State=Frozen2025.08 | 47.25 | 27.25 | 19.75 | |
| LLaVA-OVModel Family=LLaVA-OV, Model Size=<7B (0.5B), Evaluation State=Frozen2025.08 | 3.25 | 20 | 0.25 | |
| LLaVA-OV + FOCUSModel Family=LLaVA-OV, Model Size=<7B (0.5B), Evaluation State=Frozen2025.08 | 3.25 | 19.25 | 0.75 |