OpenAi has presented GPT-4.1their new family of artificial intelligence models designed for developers. This family is formed by three variants —GPT 4.1, GPT 4.1 mini and GPT 4.1 Nano—are already available through the application programming interface (API), have broader context windows and allow a better use of this context thanks to an improvement in compression.
These new models are better than the most used model of OpenAI, GPT-4O, and that its largest and most powerful model, GPT-4.5, in some aspects, according to a live broadcast Kevin Weil, Openai product director. In addition, they are “much more reliable than GPT-4o when detecting relevant text.”
But what do these AI models stand out? The three variants stand out in programming and instructions monitoring tasksthey can analyze eight times more code at the same time, improve their ability to make improvements and correct errors, They stand out in understanding of long and multimodal contexts and contextsand even have a context window of up to one million Tokens —The limit of 128,000 Tokens of the GPT-4o model.
With all these mentioned novelties, Openai seeks to enhance their products against the growing competition of companies such as Google and Anthropic, which offer similar artificial intelligence tools. In addition, Sam Altman’s company intends excel in the scope of programming with these new models.
On the other hand, with respect to the prices of these new models, This is what every AI costs:
- GPT-4.1 Nano: 0.10 cents for one million input tokens and 0.40 cents per million tokens out.
- GPT-4.1 Mini: 0.40 cents for one million entrance tokens and $ 1.60 for a million output tokens.
- GPT-4.1: 2 dollars for one million entrance tokens and $ 8 for one million departure tokens.
GPT-4.1, GPT-4O and GPT-4o Mini They are already available for integration through the OpenAi API. In addition, in response to the launch of these three models, Openai states in a release that will begin to “Discontinue the preliminary version of GPT -4.5 in the APIsince GPT -4.1 offers improved or similar performance in many key functions with much lower cost and latency. The preliminary version of GPT -4.5 will be deactivated within three months, on July 14, 2025, to give time the developers for the transition. “
Sign up to us Newsletter And receive the latest news about technology in your mail.
