:quality(85)/cloudfront-us-east-1.images.arcpublishing.com/infobae/ZF4W6XPQTREEDLSOHWACLCV4DA.jpg "Almost two decades have passed: the mysterious life of a woman and her unexpected capture in Alabama")
The company expects the next iteration of the model to be like a “PhD student” on challenging standard tasks in physics, chemistry and biology.
The ability of AI to respond to all kinds of user requests is truly amazing. But if there is one thing that is even more impressive, it is the speed with which current models respond to what is asked of them.
But it is true that some things require more thought, and New OpenAI o1 models By OpenAI It was announced Surprisingly, they are models capable of thinking to solve the most complex problems in fields such as mathematics or programming.
«We’ve developed a new family of AI models designed to spend more time thinking before responding. They can think through complex tasks and solve more difficult problems than previous models in science, programming, and mathematics.“The company announced in a statement.
These new models were announced by the company on September 12. Although they are in the “preview” or testing phase, the announcement makes the company’s most advanced and rumored AI official, which was leaked to the media a few months ago under the name “Project Strawberry.”
According to the announcement, compared to GPT-4o, OpenAI’s most advanced model, the evidence is pretty damning. In the International Mathematical Olympiad qualifying exam, GPT-4o was able to solve only 13% of the problems correctly, while the reasoning model Open Eye O1 He got 83%.
Similarly, the programming skills of this new model were evaluated in competitions and reached the 89th percentile in Codeforces competitions.
These new developments in AI with models that can interpret and analyze more complex problems in greater depth open the door to a future of general artificial intelligence, which will be more on par with human capabilities.
The company stressed that this is just a preview and hopes that upcoming updates and regular improvements will show the greater potential of these models, which are already under testing in ChatGPT and the OpenAI API.
“Social media evangelist. Student. Reader. Troublemaker. Typical introvert.”
More Stories
More than 60 Nobel Prize-winning scientists have warned of a “brain drain” of Argentine researchers to other countries.
Baringi at the Regional Science Fair: “Youth are the driving force of the coming changes”
Eight-year CNA accredited UACh Doctor of Veterinary Science