OpenAI plans to phase out GPT-4.5, its largest-ever AI model, from its API

On Monday, OpenAI announced that it will soon discontinue the availability of GPT-4.5 through its API. This massive AI model, which was only released at the end of February, will no longer be accessible to developers after July 14. They'll need to switch to another model from OpenAI's lineup, with the newly launched GPT-4.1 being recommended as the go-to alternative.

According to an OpenAI spokesperson who spoke to TechCrunch via email, "GPT-4.1 offers similar or improved performance compared to GPT-4.5 in key areas, but at a much lower cost." The company also stated its intention to phase out GPT-4.5 in order to focus on developing future models.

It's important to note that while GPT-4.5 will be removed from the API, it will still be accessible in ChatGPT for paying customers in a research preview mode. The model, code-named Orion, was trained with unprecedented amounts of computing power and data. It surpasses its predecessor, GPT-4o, particularly in writing and persuasiveness. However, despite its impressive scale, GPT-4.5 doesn't quite reach the "frontier level" on various industry benchmarks.

OpenAI has also acknowledged that running GPT-4.5 is incredibly costly. Back in February, they hinted at potential long-term limitations on its availability through the API due to these costs. The pricing for GPT-4.5 reflects this, coming in at $75 per million input tokens (about 750,000 words) and $150 per million output tokens, making it one of the priciest options in OpenAI's suite of models.