# Image Classification on ImageNet-1k (Zero-shot)

Source: [Image Classification on ImageNet-1k (Zero-shot)](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-zero-shot>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:09:06.068Z

Evaluation of Image Classification performance on the ImageNet-1k dataset using Zero-shot evaluation protocols, reporting Zero-shot Accuracy.

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 | Image Classification |
| Dataset (including any stated split) | ImageNet-1k |
| Metric | Zero-shot Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 10 of 10 |
| 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

- [Zero-shot Accuracy](<https://www.sota2.com/research/sota/image-classification-on-imagenet-1k-zero-shot.md?metric=Zero-shot%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Zero-shot Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| DFN ViT-H-14 | 84.4 | Not recorded | 68fa72f81e209131cea59a64 |
| DFN ViT-H-14 | 83.4 | Not recorded | 68fa72f81e209131cea59a64 |
| SigLIP ViT-SO-14 | 83.1 | Not recorded | 68fa72f81e209131cea59a64 |
| SigLIP ViT-SO-14 | 82 | Not recorded | 68fa72f81e209131cea59a64 |
| EVA02 ViT-bigE-14+ | 82 | Not recorded | 68fa72f81e209131cea59a64 |
| MetaCLIP ViT-H-14 | 80.6 | Not recorded | 68fa72f81e209131cea59a64 |
| MetaCLIP ViT-L-14 | 79.2 | Not recorded | 68fa72f81e209131cea59a64 |
| EVA01 ViT-g-14 | 78.5 | Not recorded | 68fa72f81e209131cea59a64 |
| OpenAI ViT-L-14 | 76.6 | Not recorded | 68fa72f81e209131cea59a64 |
| OpenAI ViT-L-14 | 75.5 | Not recorded | 68fa72f81e209131cea59a64 |

### DFN ViT-H-14

Zero-shot Accuracy: 84.4. Result ID: 6992a6033a6c31484910633c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71221e209131cea4f63c.md>)

Recorded evaluation configuration: {"Image Size":"378^2","Params (M)":986.7}. This record does not assert independent reproduction.

### DFN ViT-H-14

Zero-shot Accuracy: 83.4. Result ID: 6992a6033a6c31484910633b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa71221e209131cea4f63c.md>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":986.1}. This record does not assert independent reproduction.

### SigLIP ViT-SO-14

Zero-shot Accuracy: 83.1. Result ID: 6992a6033a6c31484910633a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ae1e209131cea4c7a4.md>)

Recorded evaluation configuration: {"Image Size":"384^2","Params (M)":878}. This record does not assert independent reproduction.

### SigLIP ViT-SO-14

Zero-shot Accuracy: 82. Result ID: 6992a6033a6c314849106339.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70ae1e209131cea4c7a4.md>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":877.4}. This record does not assert independent reproduction.

### EVA02 ViT-bigE-14+

Zero-shot Accuracy: 82. Result ID: 6992a6033a6c314849106340.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ede1e209131cea42844.md>)

[Implementation](<https://github.com/baaivision/EVA/CLIP>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":5044.9}. This record does not assert independent reproduction.

### MetaCLIP ViT-H-14

Zero-shot Accuracy: 80.6. Result ID: 6992a6033a6c31484910633e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76721e209131cea6d0be.md>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":986.1}. This record does not assert independent reproduction.

### MetaCLIP ViT-L-14

Zero-shot Accuracy: 79.2. Result ID: 6992a6033a6c31484910633d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa76721e209131cea6d0be.md>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":427.6}. This record does not assert independent reproduction.

### EVA01 ViT-g-14

Zero-shot Accuracy: 78.5. Result ID: 6992a6033a6c31484910633f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68171e209131cea1a249.md>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":1136.4}. This record does not assert independent reproduction.

### OpenAI ViT-L-14

Zero-shot Accuracy: 76.6. Result ID: 6992a6033a6c314849106338.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Image Size":"336^2","Params (M)":427.9}. This record does not assert independent reproduction.

### OpenAI ViT-L-14

Zero-shot Accuracy: 75.5. Result ID: 6992a6033a6c314849106337.

[Reporting paper](<https://www.sota2.com/research/paper/68fa72f81e209131cea59a64.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Image Size":"224^2","Params (M)":427.6}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.