Just a curiosity over whether the functionality that powers ChatGPT Agent’s web browsing will be coming to the API at some point?
I know we have the computer using agent model but that is a preview model from March and I’m assuming whatever is powering ChatGPT Agent is more powerful.
@OpenAIDevs just posted on their X account about their Agents SDK. Seems like you can spin up a multi-step agent with browsing and other tools over the API (same primitives ChatGPT Agent uses).
It’s not the same, but might be worth a look while we wait for an official “Agent” endpoint.