06 · Language & Perception
Multimodal
Multiple data types, unified
A model handling more than one data type — text, images, audio, video — in a unified way. The clear trend: frontier models are increasingly multimodal by default rather than text-only.
Concrete example
Show a model a chart image and ask it to explain the trend in words — image in, text out.
Why it matters
The direction of travel: single models that see, hear, and read at once.