OpenAI just released two powerful AI models for complex reasoning tasks. The company announced O3, their strongest reasoning model ever, along with O4-mini, a smaller but faster alternative that performs well for its size. These new tools can process images better than earlier versions, allowing them to work with sketches and visual data during their thinking process.
Both models connect with all ChatGPT features, including image generation and web browsing capabilities. ChatGPT Plus, Pro, and Team subscribers can access these models right away if they use o3, o4-mini, or o4-mini-high settings. The company plans to offer the more advanced O3-Pro model to users within weeks.
OpenAI will gradually remove older models like o1 and o3-mini to make way for these improved versions. The company faces growing competition from Chinese startups like DeepSeek, whose progress CEO Sam Altman has publicly praised. This pressure from rivals appears to have accelerated OpenAI's development timeline, coming shortly after their GPT-4.1 announcement earlier this week.
Both models connect with all ChatGPT features, including image generation and web browsing capabilities. ChatGPT Plus, Pro, and Team subscribers can access these models right away if they use o3, o4-mini, or o4-mini-high settings. The company plans to offer the more advanced O3-Pro model to users within weeks.
OpenAI will gradually remove older models like o1 and o3-mini to make way for these improved versions. The company faces growing competition from Chinese startups like DeepSeek, whose progress CEO Sam Altman has publicly praised. This pressure from rivals appears to have accelerated OpenAI's development timeline, coming shortly after their GPT-4.1 announcement earlier this week.
