OpenAI has announced that their most powerful AI model to date, GPT-4.5, will soon be discontinued for developers using their API (Application Programming Interface). This change will take effect on July 14, 2025. GPT-4.5 was only launched in late February.
For developers who have been using GPT-4.5 in their applications, OpenAI recommends switching to their newly released model, GPT-4.1. According to OpenAI, GPT-4.1 offers similar or even better performance than GPT-4.5 in key areas, but at a much lower cost.
An OpenAI spokesperson explained that this decision allows them to prioritize the development of future, even more advanced AI models.
It’s important to understand that GPT-4.5 will still be available within ChatGPT for users who have a paid subscription and are part of the research preview program. This change only affects access through the API for developers.
GPT-4.5, internally named Orion, required more computing power and data for its training than any previous OpenAI model. While it showed improvements over its predecessor, GPT-4o, in tasks like writing and persuasion, it didn’t quite reach the highest “frontier level” on some industry tests.
OpenAI has admitted that running GPT-4.5 is very expensive. This high cost was a concern even when the model was first launched, and its pricing reflects this: $75 per million input tokens (around 750,000 words) and $150 per million output tokens. This makes it one of OpenAI’s most expensive offerings.
In simple terms, if you’re a developer using GPT-4.5 through the OpenAI API, you’ll need to move to GPT-4.1 or another suitable model before July 14, 2025. This change is driven by the high cost of running GPT-4.5 and OpenAI’s focus on building the next generation of AI. Regular ChatGPT users with a paid subscription will still be able to access GPT-4.5.