OpenAI have introduced a new model – o1. It is 5 times smarter than GPT-4o
By ForGeeks OpenAI have released their long-awaited o1 model, the company’s first reasoning-capable model that significantly outperforms its predecessors. The new model performs 5 times better at solving math problems and 6 times more efficient at writing code. O1 is trained to spend more time thinking about problems before giving an answer, mimicking the human thought process.
A new model is trained to spend more time thinking about problems before giving an answer, mimicking the human thought process.
The model has shown outstanding performance in complex problems comparable to the level of PhD students in physics, chemistry and biology. In the qualifying exam for the International Mathematical Olympiad (IMO), o1 was able to solve 83% of the problems correctly, while GPT-4o could only handle 13% of the problems. In programming competitions, the model reached the 89th percentile, demonstrating outstanding ability in creating and debugging complex code.
The o1 was able to correctly solve 83% of the problems, while GPT-4o was able to solve only 13% of the problems.
Access to o1 will begin today for ChatGPT Plus and Team users, with a usage limit of 30 messages per week. Access will soon be extended to corporate and educational users, and then to all ChatGPT users.
Access to o1 will begin today for ChatGPT Plus and Team users, with a usage limit of 30 messages per week.
