OpenAI’s upgraded o3 model can use images when reasoning

OpenAI is releasing 2 caller AI reasoning models today: o3, which the institution calls its “most almighty reasoning model,” and o4-mini, which is simply a smaller and faster exemplary that “achieves singular show for its size and cost,” according to a blog post.

The institution besides says that o3 and o4-mini volition beryllium capable to “think” with images, meaning they volition “integrate images straight into their concatenation of thought.” That could beryllium utile if you amusement the models things similar sketches oregon whiteboards. OpenAI says that the models volition besides beryllium capable to set images by zooming successful connected them oregon rotating the representation “as portion of their reasoning process.”

In addition, OpenAI is announcing that its reasoning models volition beryllium capable to usage each ChatGPT tools, including things similar web browsing and representation generation. The tools volition beryllium disposable contiguous for ChatGPT Plus, Pro, and Team users successful o3, o4-mini, and o4-mini-high, and volition travel to o3-pro successful “a fewer weeks.” (o1, o3-mini, and o3-mini-high volition beryllium phased retired from those tiers.)

Today’s announcements travel OpenAI’s uncover of its flagship GPT-4.1 AI exemplary on Monday, its successor to GPT-4o.