OpenAI just released its new AI model, OpenAI o1 (internally known as Project Strawberry/Q*). The wait is finally over. Here’s the model coding an entire video game from a prompt.
The model can think before it answers and is better at math and programming challenges. The longer it thinks, the better it does on reasoning tasks. The model also ranks in the 89th percentile on competitive programming questions (!) and correctly solved 83% in a qualifying exam for the International Mathematics Olympiad (IMO).
What’s different about OpenAI o1 is that it “thinks” to produce an internal chain-of-thought before responding to the user If you use ChatGPT, you know that long threads giving more context improve responses, so this is basically what they’re doing, but from one prompt.
It’s also rolling out in ChatGPT to ALL Plus and Team users today. No waitlist!
Benchmarks
Here’s some benchmarks of OpenAI o1 compared to GPT4-o.
Also, cool to see @OpenAI in highlighting real-world use cases of the new model on their announcement post. Too many major tech companies leave this part out, but it’s important for non-technical folks to see real impact!
First impressions: It’s not as slow as I thought, and still *feels* faster than something like Perplexity (which I use daily) It also separates steps/paragraphs into sections and writes more direct, with less filler words But I’m also definitely not pushing it to its limits yet! Will share more tests soon.
Author: Rowan Cheung