---
id: "2026-03-09-fish-audio-s2-open-source"
url: "https://postcutoff.com/e/2026-03-09-fish-audio-s2-open-source/"
as_of: "2026-11-13T13:49:00+01:00"
date: "2026-03-09"
date_precision: day
category: open-source
importance: 3
confidence: high
status: [Confirmed]
sources: 4
editor: human editor
human_review: null
version: "2026-10-13"
---

As of: 2026-11-13 13:49 CET.
Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/
Canonical page: https://postcutoff.com/e/2026-03-09-fish-audio-s2-open-source/

# Fish Audio open-sources S2

Full title: Fish Audio open-sources S2: expressive 80+ language TTS with inline emotion tags

Fish Audio released S2 (S2 Pro) on 2026-03-09 with weights, fine-tuning code and an SGLang-based production inference stack: a Dual-AR TTS on a Qwen3-4B backbone trained on 10M+ hours in ~80 languages, with free-form [bracket] emotion and paralinguistic cues and multi-speaker dialogue. It led open-weights TTS on Artificial Analysis until Breeze TTS 2 (Aug 2026). The closed follow-up S2.1 Pro (June 2026) was offered as a free API.

## Key facts
- Dual-AR: 4B time-axis + 400M depth-axis; RTF 0.195, ~100 ms TTFA
- Seed-TTS Eval WER 0.54% (zh) / 0.99% (en); EmergentTTS-Eval win rate 81.88%
- API id s2-pro, $15 per 1M UTF-8 bytes; weights under Fish Audio Research License (non-commercial)
- S2.1 Pro (2026-06-23): free API tier `s2.1-pro-free` through 2026-11-30, ~90 ms TTFA, 83 languages; weights not released

## What happened

Fish Audio shipped S2 as a complete system: weights, fine-tuning code and a serving stack compatible with LLM-inference optimizations (SGLang). Emotion is controlled inline with natural-language tags.

## Why it matters

It made open TTS with fine-grained, LLM-style prompt control and production streaming available to the public, and set the open-weights bar for most of 2026. Fish Audio's later move to a free closed API (S2.1 Pro) shows price pressure in hosted TTS.

## Your AI and this story

- GPT-6 Astra (training cutoff April 2026): in its training data
- Claude Opus 5.5 (training cutoff June 2026): in its training data
- Gemini 4 Hydrogen (training cutoff March 2026): in its training data
- Grok 4.7 (training cutoff May 2026): in its training data

## Sources

1. [Fish Audio: open-sourcing S2](https://fish.audio/blog/fish-audio-open-sources-s2/) (fish.audio, official)
2. [Fish Audio S2 Technical Report (arXiv 2603.08823)](https://arxiv.org/abs/2603.08823) (arxiv.org, paper)
3. [Hugging Face: fishaudio/s2-pro](https://huggingface.co/fishaudio/s2-pro) (huggingface.co, code)
4. [Fish Audio: S2.1 Pro free API](https://fish.audio/blog/s2-1-pro-free-api/) (fish.audio, official)

## Changes

- 2026-10-13: Linked the successor Drama 3 (released 2026-10-13 under a revenue-capped Community License; S2 Pro weights stay under the Research License) and the July 2026 seed round
- 2026-09-29 (filed): Created

## Related

- 2026-10-13: [Fish Audio releases Drama 3](https://postcutoff.com/e/2026-10-13-fish-audio-drama-3/index.md)
- 2026-08-25: [BreezeBlue releases Breeze TTS 2, the new top open-weights text-to-speech model](https://postcutoff.com/e/2026-08-25-breeze-tts-2/index.md)
- 2026-07-28: [Fish Audio raises a $52M seed round led by Coreline Ventures and Capital Today](https://postcutoff.com/e/2026-07-28-fish-audio-52m-seed/index.md)
- 2026-03-23: [Mistral releases Voxtral TTS, an open-weight 4B text-to-speech model with 3-second voice cloning](https://postcutoff.com/e/2026-03-23-mistral-voxtral-tts/index.md)
- Models: [Fish Audio S2 Pro / S2.1 Pro](https://postcutoff.com/m/fish-audio-s2/) (`fish-audio-s2`)