OpenAI has introduced o3 and o4-mini, new artificial intelligence models that can think with visual data. Announced just a few days after the introduction of GPT-4.1, these technologies stand out especially with their ability to infer from visuals.
OpenAI announces artificial intelligence models that can think with images
The company describes the o3 model as the most advanced reasoning artificial intelligence to date, while the o4-mini model promises high performance and speed despite its small size. The most important feature of these new models will be their ability to integrate visuals directly into thought chains.
Users can now analyse these models by showing diagrams, tables, drawings and other visual content. In addition, operations such as rotating and zooming images will also be included in the thinking processes of the models. In this way, complex data such as technical drawings, mathematical graphs and architectural plans can be analysed in more depth.
OpenAI states that o3 and o4-mini fully support ChatGPT tools such as web browsing and visualisation. As of today, these models will be available for ChatGPT Plus, Pro and Team users. On the other hand; the old generation o1, o3-mini and o3-mini-high models will be phased out from these packages.
The new models have considerably strengthened OpenAI’s claim in the field of multimodal artificial intelligence. The ability to reason over visual data will take AI’s capacity for analysis, problem solving and creative production one step further. So what do you think about this technology? You can easily share your opinions with us in the comments section below.
