In the bustling race of AI advancements, OpenAI has launched its latest creation, the o3-Mini model, in response to DeepSeek making waves with its R1 model.
OpenAI said the o3-Mini is poised to enhance STEM capabilities across science, math, and coding — outperforming its predecessor, the o1-Mini, with faster and more accurate responses.
For those juggling between speed and complexity, developers can select from three reasoning effort levels: low, medium, and high to meet varied demands without compromising latency or accuracy.
The model is currently available to both API users and all of chatGPT users, including ones using free tier. Those on ChatGPT plus are also able to access an upper-end version of the model called “o3-mini-high,” which according to Sam Altman “thinks harder and gives better answers.”
Comparing directly with OpenAI’s o1, the o3-Mini offers a more tailored approach to technical domains, while o1 remains the general knowledge model.
Remarkably, the o3-Mini supposedly performs on par with the bigger model of the previous generation the o1 in mathematical challenges such as AIME and significantly enhances coding accuracy on Codeforces.
Chinese company DeepSeek in January released its R1 reasoning model that outperformed the o1 on benchmarks while being trained at a fraction of the costs.
Weeks into launch, the model would take the native DeepSeek app to the top of iPhone charts in the U.S. and leading AI companies like Nvidia and Perplexity rushed to integrate the model into their products and services.
The fact that DeepSeek’s upward trajectory comes with open-source models, resonating with advocates like Twitter founder Jack Dorsey, presents another challenge for OpenAI.
The company and its leader Sam Altman are in a public (and legal) battle with the world’s richest person Elon Musk over turning OpenAI into a for-profit company.
What does it take to achieve financial independence and retire early? Fire Fast by Dzambhala helps you understand and plan it out.
Join the vibrant privacy-ensured Dzambhala community on
Want to give feedback on this story? Write to us.