Skip to main content

StepFunVideo model

Step-Video-T2V 30B

Clip lengthn/a

n/alongest single cliphigher is better

Summary

Step-Video-T2V 30B is a video model from StepFun. Step-Video-T2V 30B is a text-to-video model.

What Step-Video-T2V 30B can do

“n/a” means the provider does not state it or we have not sourced it yet. That is not a “no”.

Text to videoyesSource
Image to videon/a
Referencesn/a
Audion/a
Runs locallyyesSource
Extendn/a
Start and end framen/a
Max. resolutionn/a

Head-to-head comparisons

Show all comparisons (94)

All specifications

Architecture

Parameters
30B
Source: Parameters
Checkpoints
The official release contains `stepfun-ai/Step-Video-T2V` and a separately distilled Turbo variant; this row covers only the base weights.
Source: Checkpoints

Licence and access

Weights
Open weights
Source: Weights
Access
Local
Source: Access
Release
StepFun dates the Step-Video-T2V release to February 17, 2025.
Source: Release

Further details

Input mode
Text-to-video.
Source: Input mode
Undocumented maxima
The official README documents 204 frames and example resolutions but does not establish them as maximum duration or resolution. Both maximum fields remain null.
Source: Undocumented maxima
Weights and access
Public weights on the official Hugging Face and ModelScope repositories; the release does not establish hosted provider API access.
Source: Weights and access