(CNN) – OpenAI unveiled a new artificial intelligence model that it claims can “reason” and solve more difficult scientific, coding and mathematical problems than its predecessors.
The model, the first in a series called OpenAI o1, was unveiled as a preview version on Thursday, and the company says it expects regular updates and improvements. Little by little it will be available to the majority of ChatGPT users.
“We have trained these models to spend more time thinking about problems before responding, like a person would,” explains the maker of ChatGPT on their website. “Through training, they learn to refine their thinking process, try different strategies and recognize their mistakes.”
As examples of the power of the new models, OpenAI noted that they can be used by healthcare researchers to annotate cell sequencing data and by physicists to generate “complicated mathematical formulas needed for quantum optics.”
Noam Brown, a research scientist at the company, also highlighted the potential of new AI models. “OpenAI’s o1 thinks for seconds, but our goal is for future versions to think for hours, days, and even weeks. The inference costs will be higher,” public in X on Thursday, referring to the costs — such as higher energy bills — of using AI to make inferences from input data. “But what cost would you pay for a new cancer drug? For innovative batteries? ”She asked.
On Thursday, top White House officials and top American technology executives were set to discuss AI’s massive thirst for power. Sam Altman, CEO of OpenAI; Ruth Porat, senior executive at Google; and Dario Amodei, CEO of Anthropic, attended the meeting, a person familiar with the matter told CNN.
While technology can help solve thorny problems such as cancer and the climate crisis, it poses equally complex challenges, including how to meet the significant demand for electricity required by advanced AI systems, which could worsen global warming.
According to the company, the new OpenAI model still does not have many of the functions “that make ChatGPT useful”, such as browsing the Internet for information and uploading files and images. “But for complex reasoning tasks it represents a significant advance,” he adds.
In tests, OpenAI o1 achieves similar results to PhD students on difficult tasks in physics, chemistry and biology, according to the company. And in a qualifying exam for the International Mathematics Olympiad, the new model series solved 83% of the problems correctly.
