OpenAI plans to gradually phase out GPT-4.5, the largest AI model it has ever released, from its API.
OpenAI announced on Monday that it will soon end the availability of GPT-4.5 via the API. This is remarkably fast, as GPT-4.5 was only launched at the end of February. Developers will have access to GPT-4.5 via OpenAI’s API until July 14. After that, they will have to switch to another AI model from the company’s wide range of products. OpenAI is positioning GPT-4.1, which was announced on Monday, as the recommended replacement.
When GPT-4.1 was introduced, OpenAI already made it clear that GPT-5.0 would not be the logical successor to GPT-4.5. It is unclear when GPT-5.0 will be ready. OpenAI is now focusing first on o3, o4-mini, o4-mini-high and GPT-4.1, including the nano and mini variants, as Techzine reported earlier.
GPT-4.1 offers, as stated by an OpenAI spokesperson in an email to TechCrunch, comparable or even better performance than GPT-4.5 in important areas, and at a significantly lower cost. He adds that OpenAI is phasing out GPT-4.5 to prioritize the development of future models.
To clarify: GPT-4.5 will not disappear from ChatGPT, which will remain available for paying customers in research preview. OpenAI is only discontinuing access to GPT-4.5 via the API.
No groundbreaking level
GPT-4.5, codenamed Orion, was trained with more computing power and data than any previous OpenAI release. The model performs better than its predecessor GPT-4o in areas such as writing and persuasion, but despite the scale, according to Techcrunch, it does not reach a groundbreaking level on several industry standards.
GPT-4.5 is also very expensive to use, OpenAI admits. So expensive that in February the company already indicated it was considering whether the model would remain available in the long term via the API. The pricing reflects this. GPT-4.5 costs $75 per million input tokens (approximately 750,000 words) and $150 per million output tokens, making it one of OpenAI’s most expensive models.