OpenAI released new o1 series artificial intelligence (AI) models on Thursday. The AI firm is touting these reasoning models for their advanced capabilities in solving mathematical and complex logic-based problems. There are two models – the O1, which is available in preview, and the O1-mini. The company said that these AI models were trained to spend time thinking before reacting like humans. Specifically, it is believed to be the same AI model that was reported to be a strawberry.
OpenAI o1 series AI models released
In a blog post, the AI firm introduced new AI models with advanced reasoning capabilities. These models differ from standard generative AI because it does not process the entire prompt at once and instead solves the problem systematically, as humans do. It also allows the AI model to try different strategies and correct any possible mistakes. OpenAI highlights that these models are slower than GPT-4o models because they take extra time to think.
OpenAI o1 translates a corrupted sentence. pic.twitter.com/E37e4SOuq4
– OpenAI (@OpenAI) 12 September 2024
So, what does this mean for an average user? Users will be able to ask complex questions of AI that often require multi-level reasoning and critical evaluation. For example, a question like “Look at this series: 12, 11, 13, 12, 14, 13, … What number should come next?” What requires multi-step thinking can now be solved accurately by AI.
A man goes to the library and asks the librarian for a book. The librarian points to a specific shelf. The man thanks him and leaves without taking the book. Why?
OpenAI claimed that the O1-Preview model performs at the same level as PhD students when answering questions in the subjects of physics, chemistry and biology. The model also shows similar output in solving mathematical problems. “In the qualifying exam for the International Mathematical Olympiad (IMO), GPT-4O solved only 13 percent of the problems correctly, while the reasoning model scored 83 percent,” the post said.
OpenAI CEO Sam Altman explained in an X (formerly known as Twitter) post that the O1 models were able to score 78.3 on the PhD-level science benchmark GPQA Diamond. However, he said that the Large Language Model (LLM) still has flaws as it is a nascent version of the model. OpenAI plans to release updates to continuously improve it.
For now, the o1 series AI models will be available in preview for ChatGPT Plus and Teams users. However, there is a weekly rate limit of 30 messages for o1 and 50 messages for o1-mini. The company highlighted that these limits may be increased in the future. One reason rate limits are imposed is that the model is more expensive to run than a standard transformer-based architecture.
Eligible developers will also get a chance to use the new AI model with a rate limit of 20 requests per minute (RPM). However, developers will not be able to use it for function calling, streaming, support for system messages, and more. Additionally, ChatGPT Enterprise and Edu users will get access to the model next week.
Users on ChatGPT’s free tier will soon have access to the o1-mini AI model, but it is also expected to come with a lower rate limit than GPT-4o.
