OpenAI Cancels o3 AI Model in Favor of a ‘Unified Release’

If you were beginning to get confused by OpenAI’s spiraling naming conventions and releases of its ChatGPT elements, get in line. Even OpenAI is sick of it. “We hate the model picker as much as you do and want to return to magic unified intelligence,” Sam Altman, CEO of OpenAI, posted on X.com on February 12, 2025.

GPT-4.5, the model known internally as Orion, will be ChatGPT’s last non-chain-of-thought model, according to Altman. After that, in both ChatGPT and their API, they’ll release GPT-5 as a “system that integrates a lot of our technology, including o3,” rather than offer o3 as a standalone model, as had been planned.

back to the basics

“We want to do a better job of sharing our intended roadmap, and a much better job simplifying our product offerings. We want AI to ‘just work’ for you; we realize how complicated our model and product offerings have gotten.”

So what’s a reasoning model, and how is it different from what we know as simply ChatGPT A reasoning model is a large language model that strings together a rational chain of thought when it’s posed a question, a chain that is begun and completed before it spurts out an answer.

Reasoning models “excel in complex problem solving, coding, scientific reasoning, and multi-step planning for agentic workflows,” as OpenAI puts it.

The o3 model is the immediate follow-up to the o1 model, since OpenAI skipped o2 as a designation in order not to conflict with the British mobile network operator, O2.

“While GPT-4o is still the best option for most prompts, the o1 series may be helpful for handling complex, problem-solving tasks in domains like research, strategy, coding, math, and science,” writes OpenAI in its online documentation.

The downside of that is that it typically takes longer for o1 to spit out an answer, since it has to go through its entire chain-based reasoning when you pose it a query. Rather than split OpenAI’s offerings into GPT and o3, as it currently does with GPT-4o and o1, the two lines will be merged in future offerings.

“(A) top goal for us is to unify o-series models and GPT-series models by creating systems that can use all our tools, know when to think for a long time or not, and generally be useful for a very wide range of tasks,” Altman posted.

“The free tier of ChatGPT will get unlimited chat access to GPT-5 at the standard intelligence setting (!!), subject to abuse thresholds,” he wrote.

“Plus subscribers will be able to run GPT-5 at a higher level of intelligence, and Pro subscribers will be able to run GPT-5 at an even higher level of intelligence. These models will incorporate voice, canvas, search, deep research, and more.”

OpenAI must be feeling the heat of DeepSeek, which displaced it a few weeks ago as the top free app in the Apple App Store. Doling out unlimited chat access to GPT-5, when it comes out, may just be throwing us a bone to regain dominance, but we’ll take it nonetheless. Woof.