Inter-Factuality Detection on CN Image-Text IF
67.3ScoreGPT-5.4
Evaluation results
| Method | Links | |
|---|---|---|
| GPT-5.4Model Size=≈1.5T2026.06 | 67.3 | |
| K2.5Model Size=≈1T2026.06 | 67.3 | |
| Yuvion VL-8BProtocol=Reasoning2026.06 | 66.1 | |
| GLM-5Model Size=≈700B2026.06 | 63.4 | |
| Yuvion VL-32BProtocol=Reasoning2026.06 | 62.5 | |
| Qwen3-VL-8BProtocol=Instruct2026.06 | 62.2 | |
| Qwen3-VL-32BProtocol=Instruct2026.06 | 58.6 | |
| Qwen3-VL-8BProtocol=Thinking2026.06 | 58 | |
| Yuvion VL-8BProtocol=Instruct2026.06 | 56 | |
| Qwen3-VL-32BProtocol=Thinking2026.06 | 56 | |
| Yuvion VL-32BProtocol=Instruct2026.06 | 48.3 | |
| Qwen3.5-PlusModel Size=397B2026.06 | 47.2 |