OpenAI publishes o1 models with human-like reasoning
OpenAI is releasing a new artificial intelligence model known internally as “Strawberry” that can perform some human-like rational tasks, as it looks to stay ahead of rivals in a crowded market.
The new model, called o1, is designed to spend more time calculating answers before answering user questions, the company said in a blog post Thursday. With the model, OpenAI's tools should be able to solve multi-step problems, including complex math and coding questions.
“As an early model, it still lacks many of the features that make ChatGPT useful, such as browsing the web for information and uploading files and images,” the company said. “But it's a significant advance for complex reasoning tasks and represents a new level of AI capabilities. In view of this, we are resetting the counter to 1 and naming this series OpenAI o1.”
A preview version of the model will be available to paid Plus and Team users on Thursday through OpenAI's popular chatbot, ChatGPT. Bloomberg previously reported that the company could unveil the new model as soon as this week.
The model comes as San Francisco-based OpenAI seeks to raise billions in funding and faces heightened competition in the race to develop more sophisticated artificial intelligence systems. OpenAI isn't the only company working on such capabilities; Competitors Anthropic and Google tout “reasoning” skills with their advanced AI models.
In its blog post, OpenAI gave examples of the AI model's answers to questions on various subjects, including coding, English and math, and asked it to solve a simple crossword puzzle. In a series of posts on X, Noam Brown, a research scientist at OpenAI, said the company is now releasing the model in preview to understand how people use it and where it needs to be improved.
The experience of using OpenAI's updated AI system will be slightly different from what people have come to expect with the company's chatbot ChatGPT. Before responding to a user's prompt, the new software will pause for a few seconds while, behind the scenes and invisible to the user, it considers several related prompts and then summarizes what it thinks is the best response. This technique is sometimes referred to as “chain of thought” prompting.
OpenAI has been working on making computers perform multi-step operations for some time. For example, in May 2023, the company published a blog post and an accompanying paper about efforts to improve the ability of AI systems to solve math problems. According to the paper, the company trained a model by rewarding each correct step in the process of coming up with an answer to a problem, rather than rewarding it for generating a correct answer.
Data sheet: Stay on top of the tech business with thoughtful analysis on the biggest names in the industry.
Sign up here.
About The Author
