Factuality Detection on CN Text IF
90.8ScoreYuvion VL-32B
Evaluation Results
| Method | Links | |
|---|---|---|
| Yuvion VL-32BProtocol=Reasoning2026.06 | 90.8 | |
| Yuvion VL-8BProtocol=Instruct2026.06 | 89.6 | |
| Yuvion VL-8BProtocol=Reasoning2026.06 | 88.6 | |
| Yuvion VL-32BProtocol=Instruct2026.06 | 87.7 | |
| K2.5Model Size=≈1T2026.06 | 81.9 | |
| GLM-5Model Size=≈700B2026.06 | 79.7 | |
| Qwen3-VL-32BProtocol=Instruct2026.06 | 77.6 | |
| GPT-5.4Model Size=≈1.5T2026.06 | 77.3 | |
| Qwen3-VL-8BProtocol=Instruct2026.06 | 71.3 | |
| Qwen3-VL-32BProtocol=Thinking2026.06 | 69.4 | |
| Qwen3-VL-8BProtocol=Thinking2026.06 | 66.2 | |
| Qwen3.5-PlusModel Size=397B2026.06 | 57.3 |