OpenAI’s upgraded o3 model can use images when reasoning
1 min read
Summary
OpenAI has unveiled two new AI tools, o3 and o4-mini, devised to improve the organisation’s AI reasoning capacities, with the former being the most powerful reasoning model released by OpenAI to date.
o3 and o4-mini are equipped to integrate images into their decision-making processes, useful in situations where sketches, whiteboards or other imagery play a key role.
This functionality also extends to allowing these models to manipulate images, by zooming or rotating them for instance, as they continue to think and reason.
OpenAI has furthermore extended its ChatGPT tools to its reasoning models, with all o3 and o4-mini users now having access to web browsing functionalities and image generation among other tools.
The roll-out of the new AI tools follows OpenAI’s release of its GPT-4.1 flaghsip model.