OpenAI unveiled its latest suite of artificial intelligence models on Monday, introducing GPT-4.1 along with smaller variants GPT-4.1 Mini and Nano.
The company says the new models mark a significant leap in capabilities, particularly in coding, instruction following, and long-context comprehension.
Available exclusively via OpenAI’s application programming interface, the GPT-4.1 models outperform the company’s previously top-tier GPT-4o model. They also offer access to updated knowledge up to June 2024 and can process up to 1 million tokens — a major upgrade in contextual understanding for AI systems.
“Benchmarks are strong, but we focused on real-world utility, and developers seem very happy,” OpenAI CEO Sam Altman posted on social platform X.
According to OpenAI, GPT-4.1 delivers a 21% performance boost over GPT-4o and a 27% improvement over GPT-4.5 in coding tasks. Enhancements in understanding instructions and processing long documents make the models well-suited for powering advanced AI agents.
In a notable shift, OpenAI also announced it will discontinue the GPT-4.5 preview currently offered via API by July, citing the new models’ superior or comparable performance. The GPT-4.1 family, the company added, operates at a “much lower cost” than its predecessor, potentially broadening access for developers and startups looking to build AI applications.
OpenAI first introduced a preview of GPT-4.5 in February to select users, with plans to gradually expand access. With the arrival of GPT-4.1, the company is now aiming to streamline its offerings while maintaining high performance standards.
The new releases come as AI firms race to outdo one another in the booming generative AI market, where advancements in code generation, context memory, and language understanding are rapidly reshaping software development and enterprise tools.
