Sam Altman lays out a confusing product road map for OpenAI

Let’s take a look at OpenAI’s current product lineup:

• ChatGPT 4o

• ChatGPT 4o (with scheduled tasks)

• ChatGPT 4o-mini

• OpenAI o1

• OpenAI o3-mini

• OpenAI o3-mini-high

• GPT-4

• Deep research

• Operator

• Sora

• 1-800-CHATGPT

There’s probably more, but as you can see, it’s getting a little confusing. This is almost as confusing as OpenAI partner Microsoft’s AI product offerings.

In a post on X, OpenAI CEO Sam Altman tried to clear things up with some new details about the company’s upcoming models. He offered a mea culpa with regards to the messy situation:

“We realize how complicated our model and product offerings have gotten.”

All eyes are on the pioneering startup’s next foundational models as the industry frets about reaching a possible plateau of AI model training performance, and expectations for OpenAI are high.

Codenamed “Orion,” the next model to come out will be GPT-4.5, but it sounds like this model won’t use the “chain-of-thought” approach that has emerged as a promising way forward. Chinese DeepSeek R1 models used that technique to match OpenAI’s state-of-the-art performance in some areas.

Even though GPT-4.5 isn’t out the door yet, Altman also explained that GPT-5 will be a “a system that integrates a lot of our technology, including o3.”

And while the $200 per month all-you-can-eat ChatGPT Pro offering is losing money, Altman is already promising that free users will get “unlimited chat access to GPT-5 at the standard intelligence setting (!!), subject to abuse thresholds.”

Why pay for Pro or the $20 per month Plus plan? Plus plans will get users a “higher level of intelligence” and Pro users will get “an even higher level of intelligence.”

You can read Altman’s full post below: