Openai has officially rolled out its new GPT-4.1 series, including GPT-4.1, GPT-4.1 Mini, and GPT-4.1 Nano, which users are. The company says that new models coding, instructions the following, and long references bring remarkable upgrades. Openi wrote on its blog post, “These model boards the GPT4O and GPT4O mini forward, with the major advantage in coding and instructions,” Openai wrote on its blog post.

Access to these models on Chatgpt will only be available to users to pay. In a post shared on X (East Twitter) on 14 May, openi confirmed that its latest flagship model, GPT -4.1, is now live on chat. The announcement follows a wide launch of the GPT 4.1 family on the API platform of OpenAII a month earlier, where developers can already integrate and test three versions – full, mini and nano. However, with the latest updates, the models are now available to all Chatgpt users, except free users.

What is new in GPT-4.1?

Openai claims that GPT-4.1 improves its predecessor GPT-4o in areas such as coding and instructions. The model is designed with a large reference window, which supports up to 1 million tokens. This means that it can process and keep more information at once. It also comes with the knowledge cutoff of June 2024. The knowledge cutoff of GPT 4O is October 2023.

Openai has shared benchmarks on its official blog post, which claims that GPT-4.1 shows 21 percent full improvement on GPT-4O in software engineering works and is 10.5 percent better in the following instructions. Openai says that the model is now much better in maintaining coherent interactions in many turns, making it more effective for real -world applications such as writing aid, software development and customer aid. “The benchmarks provide valuable insight, we trained these models with focus on real -world utility. Countless collaboration and partnership with the developer community enables us to adapt these models for the tasks that are most important for their applications,” says Openai.

The mini and nano variants are the scale-down version aimed at offering high performance with low cost and delay. GPT-4.1 Mini is reported to reduce delay by about half by spending 83 percent less than GPT-4O. Nano, the lightest of the three, is the cheapest and fastest model of OpenAII and is ideal for simple functions such as autocamplit or text classification.

“These models carry forward performance at every point on the delay curve,” write Openai.

Who can use it?

Only Chatgpt Plus, Pro and Team users will be able to reach GPT-4.1. Free-tier users will not get a new model, at least for now. Instead, they will continue using the GPT-4o, which Openai states that gradually incorporates the reforms seen in the new model.

GPT-4.1 is also available through API for developers and companies, Openai with more cost-skilled and powerful alternative to previous generations. New pricing involves significant cuts: GPT-4.1 input cost starts with $ 2 per million tokens, and the nano version is available only $ 0.10 per million per million tokens. In order to make the query more economical, the quick cashing discount has been increased by 75 percent.

The launch of GPT-4.1 comes as Openai has begun to phase the earlier models. GPT-4.5 Preview, a research-focused release, was promoted to API on 14 April 2025. The GPT-4, which operates the Chatgpt Plus since the model March 2023, has already been closed. While GPT-4.1 GPT-4O is not changing inside the chatgpt, its many capabilities are being converted into GPT-4O experience. However, in search of state -of -the -art performance for users and developers, direct access to GPT -4.1 through API or membership of a chat is now the way to go.