Athira Sethu
Kochi, 13 February 2025
OpenAI has announced that it is not going to bring out o3 as a separate product in the future. Instead, the company is opting for the leaner path by amalgamating o3 into the upcoming GPT-5 model. This decision, shared by CEO Sam Altman on Wednesday to let everyone know about the endeavor which is meant to provide a more integrated AI system that will gather multiple technologies together under a single roof.
In a post on X, Altman explained that OpenAI makes AI more convenient to the user and saves time. Altman informed people that it is OpenAI’s ambition to clarify any ambiguities regarding their products and just make AI “just work” like any other device. OpenAI also declared serious difficulty with the selection of the most appropriate model in the current market, and thus, it anticipates a more simplified and unified intelligence.
Within the new version that will be out, GPT-5 will use instruments of its own and also o3 will be merged into them to bring out a series of new features. Altman claimed that the users would get a “mid-intelligence level” of GPT-5 which would be the standard version of the software, but the ChatGPT Plus and ChatGPT Pro subscribers would be entitled to a higher quality AI for a more comprehensive use. The coming years will also introduce new models to have new respective features as voice, canvas, search, and deep research.
The company has not expressly given a timeline for the GPT-5 model but according to Altman it will be live in a few months. GPT-4.5, Orion is the next one and it will be out a couple of weeks before the GPT-5 so subsequent models will be held to the policy of a “chain-of-thought” architecture, meaning that the models will function differently than before but will still include beneficial enhancements in many areas.
Though there have been delays with o3, Altman thinks GPT-5 will be a big step and focus on the reliability and capabilities of the model. OpenAI is aware of the fact that they have more and more competitors, one of which is China’s AI research lab DeepSeek which has published the R1 model.
