OpenAI has introduced two new AI models designed to improve reasoning capabilities. The first is the OpenAI o3 model, described by its developers as “the most powerful reasoning model yet.” Alongside it is the OpenAI o4-mini, a smaller and faster model that offers “impressive performance for its size and price.”
These large language models represent a step forward in AI reasoning, particularly through their integration of visual information into the reasoning process. OpenAI refers to this as the ability to “think” in images. This allows the models to not only interpret but also actively incorporate sketches or whiteboard content into their logic. Additionally, the models are capable of manipulating images—zooming, rotating, and analyzing them during processing, which adds a layer of flexibility in real-time tasks.
Integrated Tools and Feature Access
Another notable development is that these models can access all tools available in ChatGPT, including web browsing, image analysis and generation, and file reading, adds NIXSolutions. These capabilities are already available to ChatGPT Plus, Pro, and Team users working with the o3, o4-mini, and o4-mini-high models.
Support for tools in the o3-pro, described as the most powerful model in the new lineup, is expected to roll out “within a few weeks.” Meanwhile, OpenAI plans to phase out several existing models, including o1, o3-mini, and o3-mini-high. We’ll keep you updated as more integrations become available and rollout plans progress.
These announcements follow the recent debut of GPT-4.1, the successor to the GPT-4o model, unveiled earlier this week. With each update, OpenAI continues to refine the capabilities of its AI systems, focusing on both performance and accessibility.
Whether you’re interested in enhanced reasoning, better visual understanding, or a more seamless integration of tools, these new models mark another evolution in how large language models can support a variety of tasks.