Inter-Factuality Detection on EN Image-Text IF
65.6ScoreYuvion VL-8B
Evaluation results
| Method | Links | |
|---|---|---|
| Yuvion VL-8BProtocol=Reasoning2026.06 | 65.6 | |
| Yuvion VL-32BProtocol=Reasoning2026.06 | 65.2 | |
| K2.5Model Size=≈1T2026.06 | 64.9 | |
| GPT-5.4Model Size=≈1.5T2026.06 | 62.9 | |
| GLM-5Model Size=≈700B2026.06 | 62.9 | |
| Qwen3-VL-8BProtocol=Thinking2026.06 | 60.7 | |
| Qwen3-VL-32BProtocol=Instruct2026.06 | 57.6 | |
| Qwen3-VL-8BProtocol=Instruct2026.06 | 56.8 | |
| Qwen3-VL-32BProtocol=Thinking2026.06 | 56 | |
| Qwen3.5-PlusModel Size=397B2026.06 | 55.9 | |
| Yuvion VL-32BProtocol=Instruct2026.06 | 54.9 | |
| Yuvion VL-8BProtocol=Instruct2026.06 | 54.7 |