OpenAI has recently introduced several updates to the ChatGPT platform. In particular, the “Reasoning Shortcut” feature has been renamed “Think” and is now available as a separate button in the query panel. This button acts as a model switch, activating the o1 model, but is visible across all models. Users can check which model generated the response by clicking the “Regenerate” button, which displays the active model. This feature is expected to support the new o3 mini model, which is rumored to be released soon.
The “o3 mini” model represents new AI capabilities from OpenAI in the field of reasoning. It is designed as a lightweight version of the larger o3 model, providing faster response times and reduced computational load, while maintaining high performance in programming, mathematics, and science tasks. OpenAI positions it as an improvement over the o1 mini model, with enhanced simulated reasoning methods for complex problems. “o3 mini” will be available for free to all ChatGPT users, indicating a shift in OpenAI’s strategy to compete with other AI models such as Google’s Gemini 2.0 Pro and Grok 3.
In addition, OpenAI has updated its flagship GPT-4o model. The latest version extends its knowledge cutoff to June 2024, enabling it to process more current events and trends. Improvements include enhanced image upload analysis and improved performance in tasks related to mathematics and programming. The model also demonstrates increased interaction with emojis, reflecting user preferences.
