OpenAI launches GPT-4.1 series with major upgrades in coding and comprehension

OpenAI has launched a new set of artificial intelligence (AI) models called GPT-4.1, along with lighter versions GPT-4.1 Mini and GPT-4.1 Nano. These models are designed to do a better job at tasks such as writing code, following instructions, and understanding long documents or conversations.

The company said GPT-4.1 performs better than its most advanced model so far, GPT-4o, and will soon replace the older GPT-4.5 model on its developer platform.

One of the biggest updates is that GPT-4.1 can now handle up to 1 million “tokens” — the small chunks of data AI models use to read and understand text. This makes it much more powerful when working with large files like long reports or big codebases.


OpenAI said GPT-4.1 is 21% better at coding tasks than GPT-4o, and 27% better than GPT-4.5. It also follows instructions more accurately and does a better job keeping track of context in longer conversations or documents.

“These models are great at coding, instruction following, and long context. Benchmarks are strong, but we focused on real-world applications, and developers seem very happy,” OpenAI chief executive officer Sam Altman said on social media platform X.


GPT-4.1 Mini and GPT-4.1 Nano are smaller versions of the main model, designed for use on devices or apps that need faster responses or have limited computing power.

The company also said the new models are cheaper to use, and added that it will turn off the GPT-4.5 preview currently available on the API in July, as the new models offer “improved or similar performance.”

OpenAI had released GPT-4.5 as a test version in February. Now, with the launch of GPT-4.1, the company is also retiring GPT-4 from ChatGPT by April 30.