Step-Audio 2 Mini Think
StepFun publishes a separate Step-Audio-2-mini-Think repository. Its card uses the Step-Audio 2 family's audio understanding and dialogue documentation.
Research release
Keep the named checkpoint distinct
The files are available as a separate Think release. The shared family text is not sufficient evidence for a specific accuracy gain over the standard Mini model.StepFun Step-Audio-2-mini-Think model card
Compare with the standard Mini checkpoint
Practical guidance. This release is most useful when the named checkpoint itself is part of the research. Run the same audio questions through both versions. Record the prompt format and output behavior before deciding whether Think is a better fit.
Specifications & access
| Detail | Value and conditions |
|---|---|
| Status | Research release |
| Access | Downloadable |
| Model identifiers | Step-Audio-2-mini-Think |
| License | The repository labels the model Apache 2.0.StepFun Step-Audio-2-mini-Think model card |
| Evidence limit | The reviewed card does not explain a separate training recipe or hosted API identifier for Think.StepFun Step-Audio-2-mini-Think model card |
Related models
- Step-Audio 2 Mini
StepFun · StepAudio 2
A speech-dialogue model supporting sound understanding, tool calls and retrieval from text and acoustic material.
Launch date not confirmedAPI · Downloadable - Step-Audio 2 Mini Base
StepFun · StepAudio 2
A base audio-language checkpoint in the Step-Audio 2 framework before dialogue tuning.
Launch date not confirmedResearch releaseDownloadable