← All models

GPT-4o Audio

Published Publisher OpenAIgpt-4o-audio-preview

Model Overview

GPT-4o Audio is a GPT-4o preview model that supports audio input and output. This entry corresponds to the fixed version gpt-4o-audio-preview-2025-06-03, which was available through the Chat Completions API at release. (community.openai.com)

Capabilities and Uses

  • Supports text and audio input and output, as well as streaming output and function calling; it can be used to build voice interaction and voice agent applications.
  • OpenAI states that this speech-to-speech model update improved instruction-following reliability, tool-calling consistency, and interruption handling.
  • The context window is 128,000 tokens, and the maximum output is 16,384 tokens; the knowledge cutoff date is October 1, 2023. (community.openai.com)

Limitations and Status

  • Does not support image or video input, structured outputs, fine-tuning, or predicted outputs.
  • This snapshot has been shut down. The OpenAI API reference lists its scheduled shutdown date as May 7, 2026; as of November 18, 2026, the official model page marks it as shut down.
  • This model is the Chat Completions version; the Realtime API version in the announcement on the same day uses a different model ID and should not be confused with it. (developers.openai.com)

Pricing

As checked on November 18, 2026, the official OpenAI model page lists the following rates in U.S. dollars (USD), all charged per 1 million tokens. The model has been shut down, and these rates do not indicate that it is currently available for use:

Type Input Output
Text tokens USD 2.50 USD 10.00
Audio tokens USD 40.00 USD 80.00

The rates above are listed on the official model page; caching, batch processing, or other discounted rates have not been verified. (developers.openai.com)

Release Date and Official Sources