GPT Audio
OpenAI's GPT Audio is a model designed to process and generate both text and audio inputs and outputs. It features an upgraded decoder intended to produce more natural-sounding voices with improved consistency.
GPT Audio specs and pricing
- Provider: OpenAI
- Released: January 19, 2026
- Context window: 128K tokens
- Max output: 16K tokens
- Input: text, audio
- Output: text, audio
- Price: $2.50 input / $10 output per 1M tokens
- Tool calling: yes
What GPT Audio is good at
- Developing voice-based conversational interfaces
- Generating natural-sounding speech from text
- Transcribing and responding to audio recordings
To-dos to try with GPT Audio
- Convert this written script into a natural-sounding audio file
- Analyze the tone and content of this audio recording and provide a text summary
- Use tool calling to trigger an action based on the spoken instructions in this audio clip
FAQ
- What types of data can GPT Audio process? The model supports both text and audio as input, and it can generate both text and audio as output.
- Who developed GPT Audio? GPT Audio is a model released by OpenAI.
- How much does GPT Audio cost? GPT Audio costs $2.50 per million input tokens and $10 per million output tokens through OpenRouter. In todo.is a typical agent step costs about 8.4 credits ($1 = 100 credits).
- What is the context window of GPT Audio? GPT Audio can read up to 128K tokens at once and write up to 16K tokens in one reply.
- Can I use GPT Audio without coding? Yes. In todo.is you pick GPT Audio as your agent's model and write to-dos in plain words, on the web, the Android app, WhatsApp or Telegram. No API key needed: credits are included, or you can bring your own key.
JavaScript is required to use the todo.is app.