GPT-4o mini TTS
OpenAI · Audio model
GPT-4o mini TTS is OpenAI's text-to-speech model for the Speech API, built on GPT-4o mini. It turns text into speech whose tone can be guided with instructions. The current default snapshot is gpt-4o-mini-tts-2025-12-15.
Current
ReleasedMar 20, 2025
Context window—
Input → outputtext → audio
API priceText input: $0.6 per 1M tokens; Audio output: $12 per 1M tokens
Consumer access—
Best for
Low-cost, instructable text-to-speech inside apps already built on the OpenAI API.
Overview
GPT-4o mini TTS is OpenAI's text-to-speech model for the Speech API, built on GPT-4o mini. It turns text into speech whose tone can be guided with instructions. The current default snapshot is gpt-4o-mini-tts-2025-12-15.
Strengths and limits
- Low price per generated audio
- Speaking style can be steered with instructions
- Still OpenAI's only dedicated speech-generation model, refreshed with a 2025-12-15 snapshot
- Input limited to 2,000 tokens per request
- Not listed in the top 10 of the Artificial Analysis TTS arena as of October 2026