OpenAI has announced a change in its strategy for releasing new AI models. Instead of the previously announced o3 model, the company has decided to focus on developing the GPT-5 model, which will integrate numerous OpenAI technologies, including elements of o3. OpenAI CEO Sam Altman noted that this decision was made to simplify the company’s product portfolio and provide users with “universal intelligence.”
OPENAI ROADMAP UPDATE FOR GPT-4.5 and GPT-5:
— Sam Altman (@sama) February 12, 2025
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.
We hate…
The GPT-5 model is expected to be available via the ChatGPT platform and the OpenAI API. Free users will get unlimited access to GPT-5 at the “standard intelligence level,” while ChatGPT Plus subscribers will be able to use it at a “higher intelligence level,” and ChatGPT Pro subscribers — at an “even higher level.” Altman did not disclose details about what these intelligence levels mean, but noted that the model will be capable of performing a wide range of tasks.
Before launching GPT-5, the company plans to release the GPT-4.5 model, codenamed “Orion.” This model will be OpenAI’s last “non-chain” model, meaning it will not use self-verification technologies. Orion faces certain technical challenges, but according to Altman, this will not prevent further work on merging different model series.
It has also become known that OpenAI continues to monitor the development of competitors, such as the Chinese lab DeepSeek, which has attracted attention with its R1 model. Altman acknowledges that competition in the market is increasing, and this motivates OpenAI to accelerate the release of new models.
