# Inference Speed on ImageNet 1024x1024 1k

Source: [Inference Speed on ImageNet 1024x1024 1k](<https://www.sota2.com/research/sota/inference-speed-on-imagenet-1024x1024-1k>)

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

Data updated: 2026-04-29T07:01:38.359Z

Evaluation of model inference speed on the ImageNet-1k dataset using inputs resized to 1024x1024 resolution, measured in seconds per image.

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 | Inference Speed |
| Dataset (including any stated split) | ImageNet 1024x1024 1k |
| Metric | Inference Time (s/img) |
| Unit | TIME |
| Direction | Lower is better |
| Results loaded | 4 of 4 |
| 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

- [Inference Time (s/img)](<https://www.sota2.com/research/sota/inference-speed-on-imagenet-1024x1024-1k.md?metric=Inference%20Time%20(s%2Fimg)>): Lower is better; unit: TIME.

## Recorded results

| Method | Inference Time (s/img) | Result date | Report paper |
| --- | --- | --- | --- |
| VibeToken-LL | 0.017 | Not recorded | 69f1906d55a3c32af89fb946 |
| LlamaGen-Tok | 0.082 | Not recorded | 69f1906d55a3c32af89fb946 |
| VibeToken-Gen | 0.46 | Not recorded | 69f1906d55a3c32af89fb946 |
| LlamaGen | 32.79 | Not recorded | 69f1906d55a3c32af89fb946 |

### VibeToken-LL

Inference Time (s/img): 0.017. Result ID: 69f1acb147f34ce924d7a6da.

[Reporting paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

Recorded evaluation configuration: {"Inference Scope":"Tokenizer enc./dec. only"}. This record does not assert independent reproduction.

### LlamaGen-Tok

Inference Time (s/img): 0.082. Result ID: 69f1acb147f34ce924d7a6d9.

[Reporting paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738a1e209131cea5d111.md>)

[Implementation](<https://github.com/FoundationVision/LlamaGen>)

Recorded evaluation configuration: {"Inference Scope":"Tokenizer enc./dec. only"}. This record does not assert independent reproduction.

### VibeToken-Gen

Inference Time (s/img): 0.46. Result ID: 69f1acb147f34ce924d7a6dc.

[Reporting paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

Recorded evaluation configuration: {"Inference Scope":"End-to-end incl. tok. dec. + generation","Model Variant":"XXL"}. This record does not assert independent reproduction.

### LlamaGen

Inference Time (s/img): 32.79. Result ID: 69f1acb147f34ce924d7a6db.

[Reporting paper](<https://www.sota2.com/research/paper/69f1906d55a3c32af89fb946.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa738a1e209131cea5d111.md>)

[Implementation](<https://github.com/FoundationVision/LlamaGen>)

Recorded evaluation configuration: {"Inference Scope":"End-to-end incl. tok. dec. + generation","Model Variant":"XXL"}. This record does not assert independent reproduction.

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