Fish Audio Changelog - What's New

All Updates (8 releases)

• S2 is here. Yeah, It's Open-source.

  Fish S2 is a big step beyond S1, redefining expressive voice AI. You can now write any emotion cues anywhere in the text, and hear the speech flow exactly how you direct it.

  March 10, 2026Use S2 Now

• Story Studio, Upgraded

  Craft long audio effortlessly with the new Story Studio. Add pauses, manage speakers, regenerate clips, and fine-tune every moment with precision.

  December 9, 2025Create now

• Hear S1, the most natural TTS ever

  Fish Audio S1 generates lifelike voices that capture emotion, rhythm, and nuance with remarkable realism.

  November 9, 2025Use S1

• S1-mini open-source release

  S1-mini (0.5B) is now open-source, bringing emotion and tone control in a compact distilled model. It delivers the core capabilities of our flagship S1 (4B), which is available on [Fish Audio](/app/).

  June 2, 2025

• Fish Speech v1.6 is here

  v1.6 brings higher stability, emotion support, and improved multilingual performance, now available for instant use online.

  March 1, 2025Use v1.6

• v1.4.0 - the most powerful open-source TTS

  Trained on 700k+ hours of audio data, Fish v1.4 is the most powerful open-source TTS model supporting more than 8 languages.

  September 12, 2024

• v1.2.0

  Open-source release of both the v1.2 pretrain and SFT model, supporting auto-reranking for stable generation.

  July 18, 2024

• v1.0.0

  Fish's first open-source TTS model release.

  April 29, 2024