On February 12, 2025, OpenAI's CEO Sam Altman announced the cancellation of the highly anticipated AI model o3, opting instead for a more streamlined product offering.

This decision marks a shift from OpenAI's initial plan to launch o3 in early 2025, as the company seeks to unify its product approach.

Altman revealed that the forthcoming GPT-5 will integrate various technologies, including elements of o3, into OpenAI's ChatGPT platform and API.

GPT-5 is set to feature advanced capabilities such as voice, canvas, search, and deep research, aiming to create a cohesive system adaptable to a range of tasks.

Before the arrival of GPT-5, OpenAI plans to release GPT-4.5, code-named 'Orion', in the coming weeks, which will be the last of its non-chain-of-thought models.

OpenAI acknowledges that non-chain-of-thought models have limitations in areas like math and physics, which they aim to address with the introduction of reasoning models.

The company is embracing reasoning models, starting with o1, designed to fact-check themselves to enhance reliability, albeit with longer response times.

OpenAI intends to offer unlimited chat access to GPT-5 at a standard intelligence setting, implementing specific thresholds to prevent abuse.

Subscribers of ChatGPT Plus will gain access to a higher intelligence setting of GPT-5, while those with ChatGPT Pro subscriptions will receive an even more advanced version.

Altman emphasized the importance of enhancing transparency regarding OpenAI's roadmap and simplifying its complex model offerings.

The shift towards reasoning models represents OpenAI's commitment to improving the reliability of its AI systems.