OpenAI Launches ChatGPT Specialized in ‘Inference’

ChatGPT developer OpenAI’s inferring ChatGPT has been unveiled. On the 12th (local time), OpenAI unveiled a new version of ChatGPT, which has been developed with a focus on inferring capabilities. It was announced that ‘Open AI o1 (hereinafter referred to as Owon)’ was launched. In the meantime, Open AI We have been developing Owon under the code name ‘Strawberry’.

Oh Won is It is an AI model specialized in reasoning through math and coding-related tasks. According to OpenAI, when Ohwon solved the International Mathematics Olympiad (IMO) preliminary exam questions, the correct answer rate was 83%, which is overwhelmingly higher than the previous model’s correct answer rate of 13%. Owon could also help physicists create complex mathematical formulas and medical researchers support experiments, OpenAI added.

“Previous models like ChatGPT start responding immediately when you ask them a question, but this model can take some time,” said Jakub Pachoki, OpenAI’s chief scientist. “It thinks about the problem in English, analyzes it, finds angles, and comes up with the best answer.”

This new model is part of OpenAI’s efforts to develop human-level AI, known as Artificial General Intelligence (AGI). OpenAI has set five levels of AI capabilities to achieve its ultimate goal of AGI.

ChatGPT Plus and Chat Teams service subscribers and businesses can use this new technology starting today. OpenAI explained, “This is a significant advance in complex inference tasks and represents a new level of AI capability. In light of this, we reset the counter back to 1 and named this model ‘OpenAI o1.'”

<ⓒ투자가를 위한 경제콘텐츠 플랫폼, 아시아경제(www.asiae.co.kr) 무단전재 배포금지>