OpenAI has unveiled two new AI reasoning models: o3, which is touted as the company’s “most powerful reasoning model,” and o4-mini, a smaller and quicker variant praised for delivering impressive performance in proportion to its size and cost, according to an announcement on the company’s blog.
The new models are designed to “think” with images, which allows for the integration of visual inputs into their reasoning processes. This feature may prove beneficial when interacting with materials such as sketches or whiteboard drawings. Furthermore, OpenAI indicated that the models could manipulate images by zooming in or rotating them as part of their analytical reasoning.
Additionally, OpenAI announced that these reasoning models would be compatible with the full suite of ChatGPT tools, which includes functionalities like web browsing and image generation. These tools will be rolled out today for ChatGPT Plus, Pro, and Team users across o3, o4-mini, and o4-mini-high, with the o3-pro variant expected to gain access in the coming weeks. Note that o1, o3-mini, and o3-mini-high will be phased out from these service tiers.
This latest announcement arrives on the heels of OpenAI’s introduction of its flagship GPT-4.1 AI model earlier this week, marking a significant advance following GPT-4o.
