Fish AudioAudio modelActive
Live 3
Summary
Fish Audio introduced Live 3 on November 10, 2026, as a full-duplex voice model for real-time conversations. It keeps listening while speaking and is designed to handle pauses, brief acknowledgements, interruptions, and corrections in conversation. It speaks English at launch.
Live 3 is available on the Fish Agent Platform and through its API. The API deploys voice agents; a standalone model ID or separate model endpoint is not documented. Fish says enterprises can deploy the model on-premises. The promotional USD 0.035-per-minute rate covers the voice-agent bundle with speech recognition, an LLM, and voice, not Live 3 alone.
Similar models
More models of the same kind as Live 3, chosen by model line, developer, size, price and release date.
- Fish Audio S2.1 ProFish Audio · same developer
- Fish Audio S2 ProFish Audio · same developer
- Scribe v3ElevenLabs · released around the same time
- Scribe v3 LiveElevenLabs · released around the same time
- AureDeepL · released around the same time
- Nemotron-3-DiarizationNVIDIA · released around the same time
- NVIDIA Nemotron 3.5 TranscribeNVIDIA · released around the same time
- MiniMax-Music3MiniMax · released around the same time
All specifications
Architecture
- Model class
- Full-duplex voice model for real-time conversations. It keeps listening while speaking and handles pauses, brief acknowledgements, interruptions, and corrections in the conversation.
- Source: Model class
Input and output
Watermarking and provenance
AI watermarking refers to invisible signals embedded in the output. Visible marks and provenance metadata are listed separately. Unknown means there is insufficient evidence for that particular field.
- AI watermarking
- Unknown
- no source
- Watermarking technique
- Not known
- no source
Licence and access
- Access
- Available on the Fish Agent Platform and through its API. Fish also says enterprises can deploy Live 3 on-premises.
- Source: Access
- Shutdown
- None announced
- no source
API
- API model ID
- No standalone API model ID or separate model endpoint ID has been published. The documented API is for deploying voice agents.
- Source: API model ID
Further details
- Weights and license
- Fish lists on-premises deployment for enterprises, but the cited pages do not publish downloadable weights or a model license.
- Source: Weights and license
Prices
- Voice-agent price
- Bundle price, not a standalone Live 3 model rate: USD 0.035 per minute in a 50% promotion through December 31, 2026. Includes speech recognition, LLM, and voice; a premium LLM is billed separately at its token rate.Source