Introducing the Reasoning Models Substack

Today the Reasoning Models Substack is born

Last week OpenAI made a foundational change to their model, and while many people expected the next model to be called GPT-5 or GPT-5o, they called it o1.

Why o1?

For the OpenAI team, this new model update was so significant, they felt like it really was starting over in many ways, because this new model operates differently from all the models before it.

What makes it different?

o1 represents OpenAI’s move into Reasoning Models.

So what’s a reasoning model?

Here’s the description directly from OpenAI because, well, they are really the best people to tell you what a reasoning model is, not me, right?

OpenAI o1 series models are new large language models trained with reinforcement learning to perform complex reasoning. o1 models think before they answer, and can produce a long internal chain of thought before responding to the user. o1 models excel in scientific reasoning, ranking in the 89th percentile on competitive programming questions (Codeforces), placing among the top 500 students in the US in a qualifier for the USA Math Olympiad (AIME), and exceeding human PhD-level accuracy on a benchmark of physics, biology, and chemistry problems (GPQA). (OpenAI)

Along with introducing a breakthrough new modeling technique, OpenAI also made a paradigm shift, introducing two different ways to access these new models, o1-preview and o1-mini. o1-preview is like previous model releases, the full monte, everything, kitchen sink included. o1-mini is a smaller version, specifically geared towards coding tasks and coming in 80% cheaper than o1-preview so that software engineers can leverage these new models without breaking the bank.

But…Reasoning Models aren’t new

Yes, OpenAI didn’t invent the concept of a Reasoning Model, this concept has existed for over a decade. There’s a great overview of Reasoning Models in Science Direct…from 2014.

If you want a good primer on the concept of Reasoning Models - this is a great place to start. Here’s the link.

So what is this Substack all about?

With OpenAI’s release of o1 I think we’re going to see a foundational shift in AI, with companies like Anthropic, Google, Apple, and pretty much everyone else jumping on the Reasoning Models bandwagon. I want to cover this shift from the beginning, starting with week one, which is why I wanted to make sure to get this first issue out today, still (just made it!) within the first week of OpenAI releasing o1.

Every week I’ll be sharing updates on how the world of Reasoning Models is changing, growing, developing, and most importantly - how you can stay ahead of the curve.

If you’ve read this far, you’ve finished the first issue. I wanted to keep this short and sweet, a quick intro to Reasoning Models, a place you can go to read more in depth about them, and an overview of what this Substack is about. Next week we’ll go deeper but that’s it for now.

Thanks for reading and stay-tuned, there’s a lot more ahead, once again, we’re still in week one!