AI-Generated Image Detection on FakeBench
88.15AccuracyDiffAIGD
Evaluation results
| Method | Links | |
|---|---|---|
| DiffAIGDModel Category=Lightweight Detector2026.03 | 88.15 | |
| Fuzzy Decision TreeModel Category=Fuzzy Decision Tree2026.03 | 87.58 | |
| VotingModel Category=Ensemble Strategy2026.03 | 87.13 | |
| Logistic RegressionModel Category=Ensemble Strategy2026.03 | 85.87 | |
| Linear SVMModel Category=Ensemble Strategy2026.03 | 85.75 | |
| DRCTModel Category=Lightweight Detector, Base Component for Ensemble=Yes2026.03 | 85.47 | |
| Qwen3-VL-8BModel Category=Multimodal Large Language Model, Protocol=Zero-Shot, Base Component for Ensemble=Yes2026.03 | 85.43 | |
| D3Model Category=Lightweight Detector2026.03 | 85.17 | |
| DDAModel Category=Lightweight Detector2026.03 | 84.77 | |
| Random ForestModel Category=Ensemble Strategy2026.03 | 84.6 | |
| Bagging TreeModel Category=Ensemble Strategy2026.03 | 84.58 | |
| LightGBMModel Category=Ensemble Strategy2026.03 | 84.55 | |
| Qwen3-VL-32BModel Category=Multimodal Large Language Model, Protocol=Zero-Shot, Base Component for Ensemble=Yes2026.03 | 82.97 | |
| Qwen2-VL-7BModel Category=Multimodal Large Language Model, Protocol=Zero-Shot, Base Component for Ensemble=Yes2026.03 | 80.97 | |
| Qwen2.5-VL-32BModel Category=Multimodal Large Language Model, Protocol=Zero-Shot2026.03 | 79.85 | |
| Qwen2.5-VL-7BModel Category=Multimodal Large Language Model, Protocol=Zero-Shot2026.03 | 75.58 | |
| UnivFDModel Category=Lightweight Detector2026.03 | 71.65 | |
| EffortModel Category=Lightweight Detector2026.03 | 71.15 | |
| AIDEModel Category=Lightweight Detector, Base Component for Ensemble=Yes2026.03 | 65.37 | |
| NPRModel Category=Lightweight Detector, Base Component for Ensemble=Yes2026.03 | 61.37 | |
| FreqNetModel Category=Lightweight Detector2026.03 | 58.95 | |
| PatchCraftModel Category=Lightweight Detector2026.03 | 50.37 |