OpenAI’s new o1 models have been creating quite the buzz since their launch. Users have shown overwhelming interest in these advanced models, prompting OpenAI to make changes quickly. But who gets access to these models, and how much can they use?
Expanded access to o1 models—big news for free and paid users
Last week, OpenAI introduced the o1 models, including the highly anticipated o1-mini and o1-preview. The models were designed to handle mathematically complex queries with more detail and reasoning capability. But there was a problem: Users, especially free users, were unable to access these models. In response, OpenAI announced on Monday that they will increase message limits for both o1-mini and o1-preview.
Previously, o1-mini was limited to 50 messages per week for Plus and Teams users. OpenAI has now increased this limit to 50 messages per day, giving paid users much more room to experiment and work with the model. For o1-preview, the limit has increased from 30 messages per week to 50 messages per week. While o1-preview will continue to be more expensive to run, this expansion brings more flexibility to users willing to pay for the $20 per month Plus membership.
Free users are starting to get access—but what’s the catch?
Here’s the real kicker—OpenAI is now rolling out access to the o1-mini model for free users. That’s right! While paid users still enjoy higher limits, free users can now test out the capabilities of o1-mini without spending a dime. Although OpenAI hasn’t officially set the limit for free users yet, early access has begun for some. If you’re one of the lucky ones, you can simply go to the model selector toggle in ChatGPT, click on “Alpha Models,” and choose o1-mini.
But before you get too excited, there’s a small catch. Free users will likely have fewer message credits than their paid counterparts, so if you’re planning on using the model for demanding tasks, you might still want to consider upgrading. Still, it’s a great way to get a feel for the new model without having to invest upfront.
What can o1 do? An inside look at its performance
Users who tried the o1-preview model were impressed by its ability to tackle some difficult tasks. From writing a WordPress plugin to finding bugs in code, the o1 model demonstrated its power to handle tasks that require a lot of reasoning and detailed output. Even if you’re not interested in coding, o1 models are known for providing more comprehensive responses to complex prompts thanks to their improved processing time.
ChatGPT o1-mini for Free users pic.twitter.com/hwDOKuaHgD
— Virág Attila (@sanchomuzax) September 17, 2024
However, it is important to note that these models do not support scanning or media uploads, so if you are after more versatile capabilities, the GPT-4o might be a better option for general requests. So whether you’re a developer or someone looking to test OpenAI’s latest model, the expansion of o1-mini’s availability gives everyone a chance to see what it can do.
Image credits: OpenAI