Project AURIONAI Use-Cases Research & Innovation
BreezeBlue
speech synthesis

Breeze TTS 2

BreezeBlue open TTS model, the only open model in the Artificial Analysis TTS top 25.

BreezeBlueBreezeBlue
Bar chart of text-to-speech models ranked by Artificial Analysis Elo score, from the model card.
Levelmodel
Classspeech synthesis
Parameters3B
Contextnot in the notes
Released25 Aug 2026
Opennessopen weights restricted
Runs onsingle datacentre GPU, consumer GPU
Statusactive

Benchmark Results

independent
Artificial Analysis Text-to-Speech Arenagenerative media
independent
1,222
1,0001,2001,400

Score in Elo; the axis starts at 1,000, as Elo has no zero. Higher is better.

rank 10 of 96, AA Text-to-Speech arena, 96 models10 Oct 2026 artificialanalysis.ai
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

R&D Topics

Large-scale zero-shot text-to-speechproduced by

Identity

Developer
BreezeBlue
Family
not in the notes
Released
25 Aug 2026 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
not in the notes Text encoder plus backbone plus depth decoder plus codec; Qwen3-TTS tokenizer
Total parameters
3B sources
Active parameters
–
Checkpoint parameters
–
Context length
not in the notes
Input
text
Output
audio
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights restricted
Licence
BreezeBlue Research and Non-Commercial licence sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
single datacentre GPU, consumer GPU
Footprint
not in the notes
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
English, Chinese only
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Lineage

The notes record no lineage for this model.

Industry Use Cases

No industry use case links this model.

Sources

  1. BreezeBlue/Breeze-TTS-2 , Hugging Face huggingface.co https://huggingface.co/BreezeBlue/Breeze-TTS-2
  2. Text to Speech Leaderboard - Top AI Speech Models | Artificial Analysis artificialanalysis.ai https://artificialanalysis.ai/text-to-speech/leaderboard
  3. BreezeBlue (BreezeBlue) huggingface.co https://huggingface.co/BreezeBlue

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Breeze TTS 2", https://aurion.semai.id/models/breezeblue-breeze-tts-2/, records of October 2026.