OpenAI announced that it will gradually remove access to its API for developers.The recently released GPT-4.5 model, switching to the new and more efficientGPT-4.1 Modelreplace.
Although when OpenAI announced the GPT-4.5 model a few days ago, it boasted that this model could communicate with users in a more natural and warm manner. At the same time, in OpenAI's SimpleQA test results, it beat GPT-37.1o, o4-mini, and o3 with an accuracy rate of 1%.
However, for the needs of developers, GPT-4.1 will be able to operate in a more efficient form. At the same time, this model is currently only available to developers, and is divided into standard GPT-4.1, small-scale GPT-4.1 mini, and even smaller GPT-4.1 nano, allowing developers to use this model in a more cost-effective form.
GPT-4.1 can cope with better encoding processing capabilities and longer context comprehension, supports up to 100 million meta-word recognition, and knows most of the "knowledge" searched through the Internet before June 2024.
Based on the cost of the median query response, GPT-4.1 will save approximately 4% in costs compared to GPT-26o, making it more suitable for developers to use it to create various automatically generated artificial intelligence services.
Currently, GPT-4.1 can be accessed through API resources provided to developers, and the existing GPT-4.5 access resources will be discontinued after July 7 this year, but this model can still be used through ChatGPT, which is publicly available.
