🚨 OpenAI's new models are here

Read time: under 4 minutes

Welcome back, Superhuman

Here’s the thing. I’m not the type of guy to send you two emails in one day. You’re a busy person. And I respect that.

But when you signed up for this newsletter, I promised you that I’d be delivering the most important AI news to your inbox every darn day. And when the biggest AI news of the year drops in the middle of the day with no warning, well, I’ll be damned if I don’t break all the rules to keep that promise.

I’ll keep this short and sweet.

OpenAI launches new series of AI models with 'reasoning' abilities

We might be halfway through September but it’s just the start of strawberry summer. OpenAI has just released its long-rumored ‘strawberry‘ models and it looks like the wait may have been worth it.

We have two new models called o1 and o1-mini and OpenAI claims that these models “can reason through complex tasks and solve harder problems than previous models in science, coding, and math.“

For example, in a qualifying exam for the International Mathematics Olympiad, GPT-4o correctly solved 13% of problems, while the reasoning model scored an incredible 83%. It also does about as well on physics, biology, and chemistry problems as a PhD student would.

So what’s the catch? It’s kinda slow. OpenAI claims the model will take its time before responding to prompts. The new models lack some of the capabilities that GPT-4o has, like web browsing and file uploads.

When can I use it? OpenAI says that ChatGPT Plus and Team users should be able to access the model starting today by picking it in the dropdown menu at the top of their screens in ChatGPT.

And that’s it for today. We’ll be back tomorrow with more updates on the new models and how you can use them.

See you then,

Zain