OpenAI appears to be preparing for the release of GPT-4.5, as references to a “GPT-4.5 research preview” have been spotted in the ChatGPT Android app. This suggests the model may launch sooner than expected.
OpenAI currently offers several models, including GPT-4o, GPT-4o with scheduled tasks, o1, o3-mini, and o3-mini-high. OpenAI CEO Sam Altman has acknowledged that these many options confuse people, and announced plans to simplify the naming system. OpenAI is now moving towards a clearer structure with versions like GPT-4.5 and GPT-5.
Some users in the ChatGPT beta app recently saw a notification saying, “Try the GPT-4.5 research preview. Pro users now have access to our newest, largest model.” Clicking the alert doesn’t do anything yet, but it is a sign that GPT-4.5 is being tested and could roll out soon. It also looks like, at first, only Pro subscribers will get access, which currently costs $200 per month.
GPT-4.5 Preview
Credit: Bleeping Computer
Alongside this new model, OpenAI is also creating new updates to the ChatGPT web interface. AI researcher Tibor Blaho noticed that OpenAI is testing changes that move features like deep research, canvas, and image-based interactions to a new file picker. If the “deep research” mode is turned on, the input field now says ChatGPT can research in the background to create detailed reports, and it suggests providing sources, focus areas, and supporting files for better results.
ChatGPT Web Update
Credit: Tibor Blaho
OpenAI has not yet officially announced the release date for GPT-4.5 or these new features. However, it is evident that they are preparing for a significant development.
About the author:
