StepAudio 2.5 ASR

Name: StepAudio 2.5 ASR
Author: StepFun

transcriptionproprietary

4B-parameter speech recognition model with multi-token prediction. Processes ~5 minutes of audio per second of compute. Supports ITN.

Capabilities

Show unsupported (13)

Infrastructure

Streaming

Model IDstepaudio-2.5-asr

ProviderStepFun

Creatorstepfun

Licenseproprietary

Statusactive

Typetranscription

Input modalitiesspeech

Output modalitiestext

Open weightYes

Parameters—

Architecture—

Languages—

Release date—

Sourceofficial

audio transcriptionsPOST

https://api.stepfun.com/audio-transcriptions

0.15 元／小时 (CNY)

GET/v1/models/stepfun/stepaudio-2.5-asr

stepaudio-2.5-asr

4B-parameter speech recognition model with multi-token prediction. Processes ~5 minutes of audio per second of compute. Supports ITN.

Changes · 5 entries

stepaudio-2.5-asrupdateca7a400Nov 2, 2026, 03:02 AM

pricing_notes["0.022 USD per hour."]→["0.15 元／小时 (CNY)"]

tagline4B-parameter speech recognition model with multi-token pr...

stepaudio-2.5-asrupdateb2f45bfOct 29, 2026, 03:29 AM

statusdeprecated→active

pricing_notes["0.15元/小时 (CNY)."]→["0.15 元／小时 (CNY)"]

stepaudio-2.5-asrupdate40decdeSep 16, 2026, 06:38 AM

statusactive→deprecated

stepaudio-2.5-asrupdatec67ed84Jul 31, 2026, 10:28 AM

pricing_notes["0.022 USD per hour."]→["0.15元/小时 (CNY)."]

tagline4B-parameter speech recognition model with multi-token pr...

stepaudio-2.5-asrcreate3e56cf1Apr 26, 2026, 07:23 PM