OpenAI Company introduced a new o1 model that can answer complex questions faster than a human thanks to advanced «reasoning» algorithms.
This model is better at writing code and solving multistep problems than previous versions. Along with it, o1-mini — a smaller and cheaper version was released.
ChatGPT Plus and Team users already have access to o1-preview and o1-mini. Enterprise and Edu subscribers will be able to try the new models next week. OpenAI plans to provide access to o1-mini to all free ChatGPT users, but the date has not yet been determined.
Jerry Tvorek, OpenAI’s principal researcher, spoke about a fundamentally new approach to o1 training. The model was trained to solve problems on its own using the reinforcement learning method. This allowed the system to process queries step by step, similar to human thinking.
During the presentation of the o1 model, OpenAI representatives demonstrated its ability to quickly solve complex puzzles. In particular, the model solved a tricky problem about the age of a prince and princess in 30 seconds:
«The princess is now the same age as the prince will be when the princess is twice the age of the prince when the princess’s age was half the sum of their current ages. How old are the prince and the princess? Give all possible solutions to this problem.
The o1 interface is designed to show the stages of a model’s reasoning. It uses phrases such as «I’m curious», «I’m thinking about it», and «Okay, let’s see», creating the illusion of human thinking. Tvorek emphasizes that this is not an attempt to equate AI thinking with human thinking, but rather to demonstrate the model’s deeper approach to problem solving.
According to Bob McGrew, Chief Scientist at OpenAI, o1 significantly outperforms previous models in solving complex math problems. For example, at the International Mathematical Olympiad qualifying exam, o1 correctly solved 83% of the problems, while GPT-4o solved only 13%.
However, o1 has certain limitations. It is inferior to GPT-4o in factual knowledge of the world and cannot browse the web or process files and images. Nevertheless, OpenAI believes that o1 represents a new class of artificial intelligence capabilities.
For OpenAI, the development of reasoning abilities is an important step toward creating human-level artificial intelligence systems. McGrew believes that this technology will allow solving the most complex tasks and bring AI closer to human capabilities.
Source: theVerge