AI models

Fish Audio

1 makers2 families4 models0 apps

Source directory →
More filters
Fish Audio/

Fish Audio

About this model or application

Fish Audio provides hosted S2.1 speech and separately licensed S2 Pro weights.

Family & other versions4 sources

Background for the family or other versions. These sources are counted separately from evidence for this release.

Model cards & docsOfficial
Fish Audio Models Overview

Family & other versions

Production and free S2.1 tiers plus the exact Transcribe Pro header.

Evidence & scope

Original checked · Oct 6, 2026

Read original
Model cards & docsOfficial
Fish Audio Pricing & Rate Limits

Family & other versions

UTF-8 byte TTS pricing, audio-hour ASR pricing and shared concurrency.

Evidence & scope

Original checked · Oct 6, 2026

Read original
Model cards & docsOfficial
Fish Audio S2 Pro model card

Family & other versions

Official S2 Pro weights, dual-AR architecture and commercial-use restriction.

Evidence & scope

Original checked · Oct 6, 2026

Read original
Model cards & docsOfficial
Fish Audio Research License

Family & other versions

Research-weight license; commercial use needs separate permission.

Evidence & scope

Original checked · Oct 6, 2026

Read original
Model family catalog3

Current models

Official catalog
  • Fish Audio S2.1 ProAvailableAPI
    Model ID
    s2.1-pro
    Released
    Not stated
    Availability
    Available
    Input
    Text · Audio
    Output
    Audio
    Deployment / access
    API
    Capabilities
    Speech generation · Multilingual

    First-party endpoint pricing

    USD / million bytes
    Input
    15

    Input UTF-8 bytes, not characters or tokens.

    83 languages; production tier with TTFA and DPA guarantees. Reference audio supports voice selection and cloning.

  • Fish Audio S2.1 Pro FreeRestrictedAPI
    Model ID
    s2.1-pro-free
    Released
    Not stated
    Availability
    Restricted
    Input
    Text · Audio
    Output
    Audio
    Deployment / access
    API
    Capabilities
    Speech generation · Multilingual

    First-party endpoint pricing

    USD / million bytes
    Input
    0

    Development/testing free tier; subject to fair-use limits, without TTFA or DPA guarantees.

    Same S2.1 model, separate access tier rather than another foundational version.

  • Fish Audio S2 ProAvailableOpen weightsAPI
    Model ID
    fishaudio/s2-pro
    Released
    Not stated
    Availability
    Available
    Input
    Text · Audio
    Output
    Audio
    Deployment / access
    API · Open weights · Local
    Weights
    Open weights
    Capabilities
    Speech generation · Multilingual · Fine-tuning

    First-party endpoint pricing

    USD / million bytes
    Input
    15

    Hosted Fish Audio API selector s2-pro; local compute is not included.

    Research/noncommercial use is free; commercial weights use needs a separate Fish Audio license. Model card describes 4B slow-AR plus 400M fast-AR; total parameters are not equated to 4B.