OpenAI released two new AI models, the OpenAI o1 and o1-mini, the company’s smartest large language models yet. They are currently available for the ChatGPT Plus and Team members, and they can think before answering, making them even smarter and far more robust with their replies.
On September 12, 2024, OpenAI published an article showcasing their new, even smarter and more advanced “o1” AI model, as they call it. But it’s not about a single feature, “a new series of AI models designed to spend more time thinking before they respond,” as the company described them.
If you are afraid of new AI models stealing your job, you aren’t going to like what comes next. But if you’ve been using them to work more efficiently, you’re gonna love OpenAI’s 1o and 1o-mini new models. The company hasn’t just created a simple large language model, it learned from it, worked on it, and now made it smarter than ever before.
What is the OpenAI o1?
If you’re a premium member, meaning that you have access to ChatGPT Plus and Team versions, you already have access to both the OpenAI o1 and o1-mini versions. For now, there’s a limitation in their usage, including the weekly rate limits of 30 messages for o1-preview and 50 for o1-mini, but those are expected to increase in the upcoming months.
While you can keep using ChatGPT, the newest versions come with enhanced reasoning capabilities, meaning that they are particularly useful if you’re trying to solve complex problems in science, coding, math, and other similar fields – or that’s what OpenAI saying.
As an early model, it doesn’t yet have many of the features that make ChatGPT useful, like browsing the web for information and uploading files and images. For many common cases, GPT-4o will be more capable in the near term. But for complex reasoning tasks, this is a significant advancement and represents a new level of AI capability. Given this, we are resetting the counter back to 1 and naming this series OpenAI o1.
OpenAI
It’s already in a state that can be used by healthcare researchers, physicists, and developers looking at creating advanced projects, generating complicated mathematical formulas needed for quantum optics, and even building multi-step workflows.
What is the OpenAI o1-mini?
Apart from OpenAI o1, there’s also a new version called o1-mini, which is exactly what it sounds like. If you’re a company or a developer looking at implementing these large language models into your applications or services, choosing between those will probably come down to your budget.
The o1 series excels at accurately generating and debugging complex code. To offer a more efficient solution for developers, we’re also releasing OpenAI o1-mini, a faster, cheaper reasoning model that is particularly effective at coding. As a smaller model, o1-mini is 80% cheaper than o1-preview, making it a powerful, cost-effective model for applications that require reasoning but not broad world knowledge.
OpenAI
The company has described both AI models exactly as it should, so I thought there wasn’t really a point in writing it differently. It’s simple; If you’re a startup company or a developer with a limited budget, the o1-mini series would be the most preferred way of implementing ChatGPT’s io series into your project.
Larger companies with bigger budgets can use the classic version of OpenAI o1 and benefit from all of its features, without getting any restrictions (hopefully in the near future) and by providing more details, information, and requests than the o1-mini series users.
How to use the OpenAI o1 and o1-mini
You can use both the OpenAI o1 and o1-mini series right now through your ChatGPT account if you’re a ChatGPT Plus or Team member. Both are available to use right now, and there aren’t any additional steps that you’ll have to go through. If you are a free member, there isn’t a way to try the two new models.
Instead of answering, these new models spend time thinking through problems before they respond, like a real person would. This is great, because, through a lot of training, these models can learn how to refine their thinking process, try different strategies, and recognize their mistakes – or that’s what the company said.
In the video above you can see more of what OpenAI o1 can do and how it helps scientists to handle the thousands of available genes. It also shows a quick preview of OpenAI o1.
If you’re looking for anything code-related, here’s everything that has been improved and what OpenAI o1 can actually do:
When it comes to coding, there’s nothing better than o1 (currently in preview, as you can see in the video above). The model is able to think before answering, making the whole structure of the code even more efficient.
From creating websites, games, applications, and whatever that may come to your mind, OpenAI o1 can probably either create it by itself, or save you tons of hours, days, months, or even years if you’re working on big and complex projects.
After watching what it can do, it makes you wonder how long my job as a web developer will last. Thankfully, if it doesn’t replace our jobs, it will make them surely a whole lot faster than before, and that’s precious.
How good OpenAI o1 actually is?
OpenAI says that their new model performs similarly to “PhD students on challenging benchmark tasks in physics, chemistry, and biology.”
They also say that it excels in math and coding. “In a qualifying exam for the International Mathematics Olympiad (IMO), GPT-4o correctly solved only 13% of problems, while the reasoning model scored 83%,” the company said in their post.
I wanted to see just how good it performs, and you know where you find such creative content, right? Social media platforms, of course! OpenAI’s profile was the first to visit, and there are two great short videos that you can watch if you want to learn more. Here’s the first:
In the video, we can clearly see how OpenAI has upgraded its new AI models. From ChatGPT reporting that the word “strawberry” has two “r’s” in it to now saying it has three, from the simplest to the weirdest, they’ve trying to make it not just good, but impossible to make mistakes.
And here’s the second one, in which ChatGPT is being used not only to solve a smart riddle but to explain it with as much detail as possible – and it does that quite well:
Conclusion: A new area has just begun
I strongly believe that this is only the beginning of what we’ll end up seeing from ChatGPT or whatever OpenAI will call their next AI models. They’ve also been moving rapidly fast with their updates and upgrades, and as their models become smarter and smarter, who knows, maybe they will start upgrading by themselves.
It’s always when you say, “I’ve seen it all,” but you haven’t seen anything. These new AI models will help millions – if not billions – of people work faster, solve complex issues within minutes, and even build whatever they have in mind but don’t have the skills to start.
Have you used the OpenAI 1o and 1o-mini? And if so, how was your experience with the new AI models and did they help you create or solve something? If you haven’t, what’s stopping you from trying them? Let me know in the comments down below.
Uh, oh! Legal stuff! Yikes.. Yet important! Affiliate Links Disclaimer: Many of the links on my articles are affiliate links. That means if you click on one and make a purchase, I might earn a small commission, and I'll keep doing what I'm doing, reviewing everything with absolute honesty. Sounds good? Good.
