In major AI tech advancement, OpenAI CEO Sam Altman has launched the company’s latest artificial intelligence models, o3, along with a mini version dubbed as o4-mini. This series can do the “thinking with images”, which means that it can analyse a user’s diagrams, sketches, whiteboards, etc., even if they are of low quality.
This marks the second major innovation in OpenAI’s reasoning models, following the launch of its first model o1 in September. The new models can manipulate these images by zooming in, rotating, or cropping the images, with the goal of enhancing the same.
The o-series also has access to various tools, including web browsing, code execution, and image generation, allowing them to perform complex tasks more efficiently. The o3 model is described as OpenAI’s most advanced reasoning system to date, while o4-mini offers a more compact and cost-effective alternative without compromising performance.
