# Deepfake Detection on COCOFake (val) Source: [Deepfake Detection on COCOFake (val)]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:29:51.238Z Evaluation of Deepfake Detection performance on the validation split of the COCOFake dataset, reporting metrics such as Accuracy and AUC. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Deepfake Detection | | Dataset (including any stated split) | COCOFake (val) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 12 of 12 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. - [AUC](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | OpenCLIP-ViT-B/32 | 99.68 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-S | 99.28 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-T | 99.25 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-M | 99.18 | Not recorded | 68fa73ea1e209131cea5eddf | | CLIP-VIT-B/32 | 99.11 | Not recorded | 68fa73ea1e209131cea5eddf | | CLIP-ResNet50 | 99.07 | Not recorded | 68fa73ea1e209131cea5eddf | | OpenCLIP-ViT-B/32 | 97.88 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-S | 93.15 | Not recorded | 68fa73ea1e209131cea5eddf | | ResNet50 | 90.31 | Not recorded | 68fa73ea1e209131cea5eddf | | ViT-B/32 | 87.64 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-M | 84.59 | Not recorded | 68fa73ea1e209131cea5eddf | | BNext-T | 83.65 | Not recorded | 68fa73ea1e209131cea5eddf | ### OpenCLIP-ViT-B/32 Accuracy: 99.68. Result ID: 6993266c3a6c314849140eb3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"OpenCLIP-ViT-B/32","Pre-training dataset":"LAION-2B","Parameters (M)":"88.3","FLOPs (G)":"8.56"}. This record does not assert independent reproduction. ### BNext-S Accuracy: 99.28. Result ID: 6993266c3a6c314849140eb8. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-S","Backbone":"trainable","Pre-training dataset":"ImageNet","Parameters (M)":"67.1","FLOPs (G)":"1.91"}. This record does not assert independent reproduction. ### BNext-T Accuracy: 99.25. Result ID: 6993266c3a6c314849140eb7. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-T","Backbone":"trainable","Pre-training dataset":"ImageNet","Parameters (M)":"29.8","FLOPs (G)":"0.89"}. This record does not assert independent reproduction. ### BNext-M Accuracy: 99.18. Result ID: 6993266c3a6c314849140eb9. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-M","Backbone":"trainable","Pre-training dataset":"ImageNet","Parameters (M)":"133","FLOPs (G)":"3.39"}. This record does not assert independent reproduction. ### CLIP-VIT-B/32 Accuracy: 99.11. Result ID: 6993266c3a6c314849140eb1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"CLIP-VIT-B/32","Pre-training dataset":"OpenAI WIT","Parameters (M)":"88.3","FLOPs (G)":"8.56"}. This record does not assert independent reproduction. ### CLIP-ResNet50 Accuracy: 99.07. Result ID: 6993266c3a6c314849140eb0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"CLIP-ResNet50","Pre-training dataset":"OpenAI WIT","Parameters (M)":"25.6","FLOPs (G)":"4.8"}. This record does not assert independent reproduction. ### OpenCLIP-ViT-B/32 Accuracy: 97.88. Result ID: 6993266c3a6c314849140eb2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"OpenCLIP-ViT-B/32","Pre-training dataset":"LAION-400M","Parameters (M)":"88.3","FLOPs (G)":"8.56"}. This record does not assert independent reproduction. ### BNext-S Accuracy: 93.15. Result ID: 6993266c3a6c314849140eb5. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-S","Backbone":"frozen","Pre-training dataset":"ImageNet","Parameters (M)":"67.1","FLOPs (G)":"1.91"}. This record does not assert independent reproduction. ### ResNet50 Accuracy: 90.31. Result ID: 6993266c3a6c314849140eae. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"ResNet50","Pre-training dataset":"ImageNet","Parameters (M)":"25.6","FLOPs (G)":"4.8"}. This record does not assert independent reproduction. ### ViT-B/32 Accuracy: 87.64. Result ID: 6993266c3a6c314849140eaf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model":"ViT-B/32","Pre-training dataset":"ImageNet","Parameters (M)":"88.3","FLOPs (G)":"8.56"}. This record does not assert independent reproduction. ### BNext-M Accuracy: 84.59. Result ID: 6993266c3a6c314849140eb6. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-M","Backbone":"frozen","Pre-training dataset":"ImageNet","Parameters (M)":"133","FLOPs (G)":"3.39"}. This record does not assert independent reproduction. ### BNext-T Accuracy: 83.65. Result ID: 6993266c3a6c314849140eb4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model":"BNext-T","Backbone":"frozen","Pre-training dataset":"ImageNet","Parameters (M)":"29.8","FLOPs (G)":"0.89"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.