'ZDNET Recommends': What exactly does it mean?
ZDNET's recommendations are based on many hours of testing, research, and comparison shopping. We gather data from the best available sources, including vendor and retailer listings as well as other relevant and independent reviews sites. And we pore over customer reviews to find out what matters to real people who already own and use the products and services we’re assessing.
When you click through from our site to a retailer and buy a product or service, we may earn affiliate commissions. This helps support our work, but does not affect what we cover or how, and it does not affect the price you pay. Neither ZDNET nor the author are compensated for these independent reviews. Indeed, we follow strict guidelines that ensure our editorial content is never influenced by advertisers.
ZDNET's editorial team writes on behalf of you, our reader. Our goal is to deliver the most accurate information and the most knowledgeable advice possible in order to help you make smarter buying decisions on tech gear and a wide array of products and services. Our editors thoroughly review and fact-check every article to ensure that our content meets the highest standards. If we have made an error or published misleading information, we will correct or clarify the article. If you see inaccuracies in our content, please report the mistake via this form.
What are o1 and o1-mini? OpenAI's mystery AI models are finally here
If you thought the recent GPT-4o and GPT-4o mini releases were enough to keep OpenAI busy, think again. Just four months after releasing GPT-4o, OpenAI finally released its rumored "Strawberry" LLM, officially known as o1 and o1-mini.
Also: 6 ways to write better ChatGPT prompts - and get the results you want faster
On Thursday, OpenAI released the o1 family of models that excel at reasoning and can solve complex math, science, and coding problems. The release came after rumors circulated for over a year that OpenAI would release an LLM that would excel at reasoning, codenamed Project Strawberry.
OpenAI CEO Sam Altman even fed into the rumors, posting a photo of a strawberry garden on X in August, accompanied by the cryptic caption, "I love summer in the garden," as seen below.
i love summer in the garden pic.twitter.com/Ter5Z5nFMc
— Sam Altman (@sama) August 7, 2024
So now that Project Strawberry AKA o1-preview and o1-mini are here, here is everything we know about the models, how you can access them, and what you can expect.
What exactly is Project Strawberry?
In May, Reuters obtained an internal OpenAI document that showed that OpenAI is working on a project codenamed Strawberry. In that internal document, Strawberry was said to be a model with advanced reasoning capabilities that could be used for deep research by navigating the internet autonomously and reliably.
Also: OpenAI trained its new o1 AI models to think before they speak - how to access them
This project is the same as Q*, a model whose rumors began circulating in November 2023. At that time, Reuters reported that some OpenAI employees regarded Q* as a breakthrough in artificial general intelligence (AGI).
The official answer to these rumors is the o1-preview and o1-mini models, which OpenAI released on September 12. Unlike the rumors above, the models are not AGI or fully autonomous, but they are indeed capable of advanced reasoning that researchers and scholars can leverage. The models are already available in ChatGPT for ChatGPT Plus and Team subscribers.
What can you use o1-preview and o1-mini for?
The major selling point of both models is that they excel at reasoning, AKA thinking questions through. OpenAI shared that it trained the models to think before responding as humans do, which refines their thinking process and allows them to try different strategies and identify their mistakes.
Also: Want Apple's new AI features without buying a new iPhone? Try this app
Because they can reason through prompts, they perform better on hard or complex math, science, and coding problems. Therefore, when deciphering what model to use, the best course of action would be to opt for o1-preview when requiring assistance with more complex problems. OpenAI shares o1-mini is particularly effective at coding.
How do the o1 models compare to GPT-4o?
The competitive edge of the o1 models is their ability to excel at reasoning. This is due to the training approach that taught them to think before responding and, as a result, allowed them to try different strategies before landing on an answer.
Therefore, the o1 model is very strong in math, science, and coding. For example, it scored 83% on the International Mathematics Olympiad (IMO) qualifying exam. For comparison, GPT-4o correctly solved only 13% of problems. As a result, when solving a hard or complex STEM problem, this model is your best option.
Also: 10 features Apple Intelligence needs to actually compete with OpenAI and Google
Even though this performance might lead you to believe it is a better alternative in all case scenarios, that is not the case. Because o1 is an early model, it lacks key ChatGPT features, such as internet browsing and accepting media uploads, and as a result, for now, GPT-4o remains the better option for most requests.
ChatGPT Plus and Teams users can toggle between the different models depending on the prompt, so they don't have to commit to using one or the other. To make it easier to decide which to use, OpenAI is working on a feature that allows ChatGPT to choose the best model automatically based on user prompts.
Where can Project Strawberry/o1-preview and o1-mini be accessed?
ChatGPT Plus and Team users can access the o1-preview and o1-mini models from the model picker toggle on the upper-left hand side of their ChatGPT page, with weekly rate limits of 30 messages for o1-preview and 50 for o1-mini.
Also: What does GPT stand for? Understanding GPT-3.5, GPT-4, GPT-4o, and more
Altman confirmed the rollout was live to all ChatGPT Plus/team users. Eventually, OpenAI plans to bring o1-mini to all ChatGPT free users, too, but did not explicitly say when that change will happen.
Of course, developers also have the opportunity to access the models, specifically, those who qualify for API usage tier 5 in the API with a limit of 20 RPM. ChatGPT Enterprise and Edu users will get access at the beginning of next week.
Is o1-preview and o1-mini free?
These models are not yet free, but OpenAI plans to make them available to everyone. Until then, the only way to access o1-preview and o1-mini is through one of the paid subscriptions, including ChatGPT Plus, Teams, Enterprise, and Edu. Developers can also access through the API.
Also: ChatGPT vs. ChatGPT Plus: Is a paid subscription still worth it?
ChatGPT Plus is the best way for individual users to access the models; it costs $20 per user per month and comes with other perks, such as priority access to the latest models and the ability to create GPTs.