There is OpenAI opened its latest AI model, the o1, which the company claims can perform complex reasoning tasks more effectively than its predecessors. The release comes as OpenAI faces increasing competition in the race to develop more advanced AI systems.
OpenAI said on its website that O1 was trained to “spend more time thinking about problems before responding, like a human.” “Through training [the models] learn to improve their thought process, try different strategies and recognize their mistakes.” OpenAI envisions the new model being used by health researchers to interpret cell sequence data, by physicists to create mathematical formulas and by software developers.
Current AI systems are essentially fancier versions of autocomplete, generating answers through statistics instead of actually “thinking” through the question, which means they’re less “intelligent” than they seem. Engadget time tried to get ChatGPT and other AI chatbots to solve New York Times For example, in the Spelling Bee, they made mistakes and produced meaningless results.
With O1, the company claims it’s actually “setting the counter back to 1” with a new kind of AI model designed to handle complex problem solving and logical thinking. a blog post detailing the new model, OpenAI said it performs like PhD students on challenging physics, chemistry and biology tasks, and excels at math and coding. For example, its current flagship model, the GPT-4o, solved only 13 percent of the problems on the International Mathematical Olympiad qualifying exam, while the o1 solved 83 percent.
However, the new model does not include capabilities such as browsing the web or downloading files and images. And, according to The VergeSignificantly slower in processing queries compared to GPT-4o. Despite having more time to review its results, o1 hasn’t solved the problem of “hallucinations” — a term for AI models that organize information. “We can’t say we’ve solved hallucinations,” said Bob McGrew, the company’s chief research officer. The Verge.
O1 is still in its infancy. OpenAI calls it a “preview” and is making it available only to ChatGPT customers for a fee starting today, with limits on how many questions they can ask per week. In addition, OpenAI also offers o1-mini, a slimmed-down version that the company says is particularly effective for coding.
