Step1X-Edit
Step1X-Edit combines a multimodal language model that interprets edits with a diffusion transformer that generates the image. StepFun published the original weights on 25 April 2025.
Interpret the edit before rendering it
The visual-language component produces editing tokens from the instruction and image. A DiT network decodes them into the edited result.StepFun Step1X-Edit model card
Choose weights deliberately
Practical guidance. This release remains relevant for local inference and reproducible editing work. An old hosted API example no longer establishes service access. Compare the later 1.2 model when an edit needs reasoning about an abstract instruction.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | Downloadable |
| Model identifiers | Step1X-Editstep-1x-edit |
| License | Apache 2.0.StepFun Step1X-Edit model card |
| Hosted access | The old step-1x-edit service stopped before the wider image API retirement on 10 October 2026.StepFun image API retirement notice |
Related models
- Step1X-Edit 1.2
StepFun · Step1X
An image-editing model that adds reasoning and reflection to the Step1X-Edit pipeline.
Launch date Downloadable - Step1X-Edit 1.1
StepFun · Step1X
An instruction-based image-editing model that also supports text-to-image generation.
Launch date Downloadable