GPT-4o
AI Model
Multimodal
Prior generation
'Omni': one network for text, vision and real-time speech.
You're looking at an AI model record: who built it, what it can do, and the hardware and data behind it.
GPT-4o merged modalities into a single end-to-end model — real-time voice conversation with ~300 ms latency, native image understanding and generation hooks — and became ChatGPT's free default in 2024, the moment multimodal stopped being a demo.
How this connects to the rest of the stack
This record's links at a glance — click a group to jump to it.
Developed by
Succeeded by
Successor to
Competes with
Related
Sources
Record reviewed as of 2026-01-15 — see Methodology.