OpenAI has begun making the new GPT-4.1 AI model available to all subscribers of paid ChatGPT plans.
Included in the release is also a lighter version known as GPT-4.1 mini, which will be accessible to all users free of charge.
Phased Rollout Includes All Paid Tiers
According to the official announcement, the new model started reaching users on Plus, Pro, and Team plans, while Enterprise and Edu subscribers will receive the update in the coming weeks.
Users are able to select GPT-4.1 from the list of available models within the ChatGPT platform.
Notably, GPT-4.1 arrives shortly after the launch of GPT-4.5 last February, reflecting the accelerated pace of development within OpenAI.
The company indicated that the new model outperforms the previous version in terms of speed and execution efficiency, especially in coding tasks and instruction following.
New Free Version Named GPT-4.1 mini
In addition to the full version, OpenAI announced the launch of GPT-4.1 mini to serve as the official alternative to GPT-4o mini.
This update will reach all users, including those using the free version of ChatGPT.
Furthermore, GPT-4.1 mini will automatically be set as a fallback model when the usage limit for GPT-4o is exceeded.
New Hub for Safety Evaluations and Transparency
In this context, OpenAI unveiled a new safety evaluations hub on its official website. Its purpose is to provide greater transparency regarding model performance concerning potential "hallucinations," instruction adherence, and generating harmful content.
The hub displays continuous updates on model safety and is an extension of the technical card system the company previously used upon releasing each version.
The company affirmed in its statement that this direction is part of its effort to enhance trust with the technical and research communities. It stated: "We aim to share safety evaluation results regularly to support community efforts that drive towards greater clarity in the field of AI."
Controversy Over Transparency
Despite this initiative, some experts expressed reservations about OpenAI being the sole party conducting these evaluations and controlling what is published to the public. This raises questions about the completeness of the picture presented by the company.
Previously, OpenAI faced criticism for releasing GPT-4.1 without a detailed safety report, leading some researchers to accuse the company of reducing transparency standards.
The company responded by stating that GPT-4.1 is not classified as a "frontier" model requiring higher safety standards, as it does not surpass the 'o' series in its capabilities.
Increasing Competition in Coding Tools
The launch of GPT-4.1 coincides with a rising wave of interest in AI-powered coding tools.
Reports indicate that OpenAI is nearing the acquisition of Windsurf, one of the most prominent AI-supported coding tools.
Conversely, Google announced a new update to its Gemini service, aimed at facilitating its connection with GitHub projects.
