Skip to content

FireRedASR

An open industrial Mandarin speech recognition model

Visit official site

What FireRedASR is

The model is described as an open industrial-grade Mandarin ASR model built on an Encoder-Decoder and LLM architecture.

What you can do with it

  • Handle Mandarin sources such as video and livestream
  • Integrate the model into an application

Pros & cons

✓ What we like

  • Open model and inference code are provided
  • Dialect and English recognition are listed

! What to watch out for

  • The focus is Mandarin speech recognition
  • A variant must be chosen by task need

FAQ

What variants exist?

The material names FireRedASR-LLM and FireRedASR-AED.

Is it open?

Open models and inference code are described for developers.

Last reviewed: 2026-11-15

More AI audio tools tools

View all →

How we review