OpenAI has introduced new AI models called the Strawberry series. These new models can think longer and solve hard problems better than the older ones. They are especially good at answering difficult questions in subjects like science, coding and math.
Inside OpenAI the project was called Strawberry but the models are named “o1” and “o1-mini” and the “o1” model is now available for ChatGPT and its API.
Noam Brown who is a researcher at OpenAI posted on social media that these models are part of the Strawberry project. He said he is excited about these new AI models because they can think and reason like a human.
The new “o1” model did very well on a math test getting a score of 83%. The older model GPT 4o only scored 13%. The “o1” model also did better than people with PhDs on programming and science problems
Reuters first reported on this project which was originally called “Q*” in November 2023. Later it was named “Strawberry.”
