Reasoning on FakeClue++
95.1Rel ScoreFakeVLM-R1
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FakeVLM-R12026.05 | 95.1 | 91.9 | 77.7 | |
| Gemini-2.52026.05 | 93.6 | 90.4 | 76.5 | |
| GPT-52026.05 | 93.1 | 84.3 | 67.2 | |
| GPT-4o2026.05 | 92.8 | 81.4 | 50.8 | |
| Qwen2.5-VLParameters=32B2026.05 | 92.2 | 84.4 | 44.5 | |
| Qwen2.5-VLParameters=7B2026.05 | 91.3 | 84.3 | 44.4 | |
| Skyra2026.05 | 84.9 | 79.5 | 53.5 | |
| DeepSeekVL2Parameters=7B2026.05 | 77.9 | 72.3 | 33.5 | |
| BusterX++2026.05 | 74.1 | 69 | 46.5 | |
| SIDA2026.05 | 73.1 | 61.6 | 40 | |
| InternVL3Parameters=38B2026.05 | 70.4 | 63.1 | 35.7 | |
| InternVL3Parameters=8B2026.05 | 69.4 | 62.6 | 36.5 | |
| FakeVLM2026.05 | 31 | 20.4 | 13.1 |