OpenAI’s o1 and o1-mini AI models on ChatGPT have solved word, math and chemistry puzzles
OpenAI released a preview of its new AI model series, called o1 and o1-mini and nicknamed Strawberry, on Thursday, which the tech company claims can answer complex science, math and coding questions as it is capable of performing sophisticated, humanlike reasoning.
ChatGPT Plus and Team users can access the model.
The technology aims to help programmers, physicists and researchers by reasoning through tasks step-by-step. Unlike before, this reasoning process is now automated, enabling the models to think through problems independently.
Superior problem-solving capabilities
The o1 model, available in ChatGPT and via API, scored 83% on the International Mathematics Olympiad’s qualifying exam, a major leap from GPT-4o’s 13%.
The chatbot also solved a complex acrostic puzzle and answered advanced chemistry questions.
OpenAI is going big
The company is raising funds from investors valued at $250 billion.
However, it faces several copyright and intellectual property challenges, including a pause on GPT-4o’s Sky voice after allegations that it mimics actress Scarlett Johansson’s voice from the movie Her. It also faces lawsuits from newspapers and publications.