OpenAI has introduced its “Strawberry” AI models, named o1 and o1-mini, designed to improve reasoning and solve complex problems. the company stated in a blog post Thursday. Available in ChatGPT and its API, the o1 model significantly outperforms previous models, scoring 83% on the International Mathematics Olympiad qualifying exam compared to 13% by GPT-4o.
The models also excel in competitive programming and science benchmarks using “chain-of-thought” reasoning, which breaks down complex problems into smaller steps. OpenAI has automated this process, allowing the models to think through problems more carefully, similar to human problem-solving.
