Together AI launches Qwen3 TTS inference service
Understand this story
Together AI has launched an inference service for the open-source text-to-speech model Qwen3 TTS (Qwen/Qwen3-TTS-12Hz-1.7B), which developers can run on its Serverless Inference and dedicated endpoints. Together AI says that under streaming responses, the first audio packet arrives in as little as 97ms, a latency based on its service stack tuned for real-time speech. The model is released by Qwen under the Apache 2.0 license, offers 9 preset voices, allows natural language instructions to control emotion and tone, and covers 10 languages including Chinese, English, and Japanese with the same voice.
AI overview based on source reports · Updated 2 hr ago
Coverage timeline
Follow the coverage to explore different perspectives.
- Together AI BlogQwen3 TTS is now available on Together AI
Together AI now serves Qwen3 TTS, an open text-to-speech model with nine preset voices, plain-language delivery instructions, and support for 10 languages, with first audio in as little as 97ms. It is available on Together Serverless Inference and dedicated endpoints, supports HTTP and WebSocket streaming, and costs $20 per 1M characters.
Activity for this story
Not enough continuous observations to show a trend yet.