High operating costs force OpenAI to stop providing its largest AI model via API, replacing it with GPT-4.1 with equivalent performance but lower costs.
OpenAI has just announced plans to soon stop providing the GPT-4.5 model, the company's largest AI model to date, through their API. This announcement was made just about two months after GPT-4.5 was launched at the end of February.
According to the announcement, developers will continue to have access to GPT-4.5 via OpenAI's API until July 14, after which they will need to switch to another model in the company's product lineup. OpenAI is positioning GPT-4.1, which was launched last Monday, as the preferred replacement.
"[GPT-4.1] provides equivalent or better performance compared to GPT-4.5 in key areas with much lower costs," an OpenAI representative shared via email with TechCrunch. "[W]e will remove GPT-4.5 to prioritize the development of future models."
To clarify, GPT-4.5 will not disappear from ChatGPT, where it is still provided as a research preview for paid customers. OpenAI is only removing it from the API.
Enormous operating costs behind the decision
GPT-4.5, with the code name Orion, was trained with computational power and data larger than any previous version of OpenAI. It improves upon its predecessor, GPT-4o, in areas such as writing and persuasive capabilities, but despite its large scale, GPT-4.5 has not yet reached the "pioneering level" on some industry evaluation standards.
OpenAI acknowledges that GPT-4.5 is very expensive to operate – so costly that the company warned in February that they were evaluating whether to provide GPT-4.5 through their API in the long term. The model's price reflects this: GPT-4.5 costs $75 per million input tokens (about 750,000 words) and $150 per million output tokens, making it one of OpenAI's most expensive products.
This decision marks a turning point in OpenAI's product development strategy, as the company seems to be weighing performance against commercial accessibility. The launch of GPT-4.1 with equivalent performance but lower costs suggests that OpenAI is seeking to optimize its product lineup while dedicating resources to developing more advanced AI models in the future.
