Azure Speech in Foundry Tools pricing

See the documentation for information on quotas, limits and instructions on how to increase concurrent requests.

Speech to text hours are measured as the hours of audio sent to the service, billed in second increments.

1To take advantage of this new Batch Transcription pricing you need to use Speech to text REST API V3.2 or later versions. See Speech to text REST API for information.

2This reflects public preview pricing.

3This price includes 1 audio input and output, up to 2 text translation language using standard or custom Speech to Text and standard Translation. For custom Translation or 3+ translation languages, please reference the Azure Translator in Foundry Tools Text Translation pricing page.

4Selected text to speech voices are available via two model variants: Neural and NeuralHD. Learn more here.

5Custom Speech Training applies when customizing any base model released on or after October 1, 2023.

6Personal Voice is a limited access feature restricted to certain pre-approved use cases only, with a need to applying for access. To learn more about the service, check the document.

7Speaker Recognition is a limited access feature with a need to apply for access.

8Text to Speech: speech synthesis usage is billed per character. Avatar is billed per second. Training and model hosting is billed per second.

9To use Fast Transcription you need to use Speech to text REST API 2024-05-15-preview or later versions. See Speech to text REST API for information.

VL1With Voice Live Pro, developers can choose from larger LLMs such as GPT-Realtime, GPT-4o and GPT-4.1 models. With Voice Live Standard, developers can choose from smaller LLMs such as GPT-4o-Mini-Realtime, GPT-4o Mini and GPT-4.1 Mini models. With Voice Live Lite, developers can choose from SLMs and equivalent models such as GPT-4.1 Nano and Phi models. Models for each tier will be updated or retired as they become available. To learn more how Voice Live API pricing works, click here.

VL2You will be charged separately for custom speech and custom voice model training and hosting. Refer to the ‘Speech to Text - Custom Transcription’ and ‘Text to Speech - Custom Voice - Professional’ pricing for details. Custom voice is a limited access feature. Learn more about how to create custom voices.

LIThis price includes text output