The company said its system, currently being tested by safety and security experts, performed better than top AI technologies in tests measuring math, science, coding, and logic skills.
San Francisco: OpenAI announced a new AI system called OpenAI O3 on Friday, which is designed to solve problems in math, science, and computer programming.
The company said it is currently sharing the system only with safety and security testers. It also said the system performed better than top AI technologies in tests that measure skills in math, science, coding, and logic.
The new system is an improved version of O1, the reasoning system launched earlier this year. OpenAI said O3 was over 20% more accurate than O1 in common programming tasks and even performed better than its chief scientist, Jakub Pachocki, in a programming competition. The company plans to make the technology available to people and businesses early next year.

Comments
Post a Comment