Visual Reasoning on V* Attribute, Spatial, Overall
83.3Overall AccuracyDeepEyes 7B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| DeepEyes 7BPrecision=BF16, Batch size=1, Model Parameter Scale=7B2026.06 | 83.3 | — | — | |
| LVR 7BPrecision=BF16, Batch size=1, Model Parameter Scale=7B2026.06 | 81.7 | 84.4 | 77.6 | |
| Qwen2.5-VL 32BPrecision=BF16, Batch size=1, Model Parameter Scale=32B2026.06 | 81.2 | 77.4 | 86.8 | |
| EVA-VL 7BPrecision=BF16, Batch size=1, Model Parameter Scale=7B2026.06 | 80.2 | 80 | 80.3 | |
| Qwen2.5-VL 7BPrecision=BF16, Batch size=1, Model Parameter Scale=7B2026.06 | 76.4 | 78.2 | 73.6 | |
| GPT-4oPrecision=BF16, Batch size=12026.06 | 67.5 | 72.2 | 60.5 |