OpenAI will gradually phase out GPT-4.5 from its API by July 14 due to high costs

On Monday, OpenAI announced the discontinuation of access to the GPT-4.5 model through its API by July 14. The GPT-4.5 model, released at the end of February, will be replaced by the GPT-4.1 model. Developers can use GPT-4.5 via OpenAI's API until the specified date.

A representative of OpenAI told TechCrunch: "GPT-4.1 offers similar or improved performance to GPT-4.5 in key areas, but at significantly lower costs." The company notes that discontinuing support for GPT-4.5 will allow them to focus on developing future models.

GPT-4.5, codenamed Orion, is trained using more computational power and data compared to previous versions of OpenAI. The model outperforms GPT-4 in writing and persuasiveness areas.

GPT-4.5 remains available for paid ChatGPT customers, but it's gradually being phased out from the API. The company acknowledges the high operational costs associated with the model: GPT-4.5 costs $75 per million input tokens and $150 per million output tokens, making it one of OpenAI's most expensive offerings.