Multi-image understanding on BLINK multi-img
55.6AccuracyQwen2.5-VL-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2.5-VL-7BParameters=7B2025.12 | 55.6 | |
| InternVL-2.5-8BParameters=8B2025.12 | 54.8 | |
| Qwen2-VL-7BParameters=7B2025.12 | 52.9 | |
| SDAR-VL-8B-InstParameters=8B, Training=Instruction-tuned, Controlled Group=true2025.12 | 52.9 | |
| SDAR-VL-4B-InstParameters=4B, Training=Instruction-tuned, Controlled Group=true2025.12 | 52.8 | |
| LLaVA-OV-7BParameters=7B2025.12 | 52.4 | |
| GPT-4VProprietary=true2025.12 | 51.1 | |
| InternVL-2.5-4BParameters=4B2025.12 | 50.7 | |
| InternVL-2-8BParameters=8B2025.12 | 50.3 | |
| Qwen2.5-VL-3BParameters=3B2025.12 | 48.8 | |
| InternVL-2-4BParameters=4B2025.12 | 47.4 |