Multimodal flagship — text, vision, audio. Default OpenAI workhorse.
No agents are using this model yet.