Linear classification on frozen ImageNet-1K features. Used to evaluate representation quality of self-supervised and contrastive models without fine-tuning the backbone.
Top1 Accuracy is the reported evaluation metric for ImageNet Linear Probe. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.
Higher is better
Muted rows were not state of the art when published — an earlier or same-year result already scored better.
| Rank | Model | Trust | Score | Year | Links | Fix |
|---|---|---|---|---|---|---|
| 01 | DINOv2 ViT-g/14 | vendor | 86.5 | 2026 | Source ↗ | Looks wrong? |
| 02 | SimCLRv2 (ResNet-152 3x) | vendor | 79.8 | 2026 | Source ↗ | Looks wrong? |
| 03 | MAE ViT-H/14 | vendor | 76.6 | 2026 | Source ↗ | Looks wrong? |
Top 1 Accuracy is the reported evaluation metric for ImageNet Linear Probe. Codesota tracks published model scores on this metric so readers can compare state-of-the-art results across sources and model families.
Higher is better
Muted rows were not state of the art when published — an earlier or same-year result already scored better.
| Rank | Model | Trust | Score | Year | Links | Fix |
|---|---|---|---|---|---|---|
| 01 | DINOv2 ViT-g/14 | verified | 86.5 | 2026 | Source ↗ | Looks wrong? |
| 02 | DINOv2 ViT-L/14 | verified | 86.3 | 2026 | Source ↗ | Looks wrong? |
| 03 | CLIP ViT-L/14 | verified | 85.3 | 2026 | Source ↗ | Looks wrong? |
| 04 | MAE ViT-H/14 | verified | 77.2 | 2026 | Source ↗ | Looks wrong? |
| 05 | MAE ViT-L/16 | verified | 76 | 2026 | Source ↗ | Looks wrong? |