GPT-4o (“o” for omni) handled text, images and audio in a single model. It was noticeably faster than GPT-4 Turbo and priced lower, and OpenAI made it available to free ChatGPT users.
AI model data sheet: GPT-4o
| Developer | OpenAI |
|---|---|
| Released | May 13, 2024 |
| Type | Natively multimodal model (text, image, audio) |
| Context window | 128K tokens |
| Access | ChatGPT (including free tier) and API |
| Licence / weights | Closed |
| Best at | Fast, cheap everyday tasks, voice and image understanding |
| Why it matters | Put a GPT-4-class model in free users’ hands |
Quick verdict: The default all-rounder of mid-2024.
Its real-time voice demo drew most of the attention at launch.