OpenAI has launched its latest models, o3 and o4-mini, to enhance the performance and speed of ChatGPT in reasoning tasks. Both models boast a remarkable ability to autonomously select from a suite of tools, including searching the web, analyzing uploaded files and other data, improved deep reasoning with visual inputs, and generating images.

For the first time, ChatGPT can integrate uploaded images into its reasoning framework, a crucial step for real-world applicability and engagement. OpenAI claims that these new models set a higher standard for intelligence and usefulness in AI, successfully tackling complex, multifaceted questions. The new reasoning models do not only yield intelligent conclusions but also detail their processes, showing each step taken and which tools were employed, thus enhancing transparency and user engagement.

The integration of multimodal reasoning, where visual inputs can be analyzed alongside text inquiries, is a remarkable leap in how AI assistants work.  This release enables ChatGPT to reason with images, allowing it to perform tasks such as identifying vehicles and other objects, from photographs and assessing their future retail values.

Users can expect a more responsive and capable tool that promotes a richer interactive experience. The improvements in detail and usability are already evident. Users report quicker and more comprehensive responses compared to previous iterations, such as o1 and o3-mini.

OpenAI says that the o3-pro model will be released soon, with enhanced tool support, suggesting that this is just the beginning of a new era in AI capabilities. The o3 and o4-Mini models are available for ChatGPT Pro, Plus, Team users. It is also available for free-tier users with some limitations. These are impressive advancements, and more could be on the horizon as AI technology continues to evolve.

Source: TechRadar