OpenAI Launches New GPT-4.1 Models

April 14, 2025 — Artificial intelligence pioneer OpenAI has officially launched its next-generation model, GPT-4.1, alongside its smaller variants — GPT-4.1 mini and GPT-4.1 nano. These models come with significant enhancements in coding capabilities, instruction following, and long context comprehension.

Available exclusively via the OpenAI API, these models are said to outperform even the advanced GPT-4o, previously considered the company’s most capable release.

The GPT-4.1 family supports up to 1 million tokens, allowing the models to process large amounts of information in one go. A “token” is a unit of data the AI uses for understanding and generating content. Additionally, the models come with updated knowledge up to June 2024, making them more relevant for developers and businesses.

According to OpenAI, GPT-4.1 outperforms GPT-4o by 21% and GPT-4.5 by 27% in coding benchmarks. Enhanced instruction adherence and the ability to handle long context make these models especially suitable for powering advanced AI agents.

OpenAI CEO Sam Altman shared on X:

“Benchmarks look great, but our focus was real-world usability — and developers are very pleased.”

The GPT-4.1 models also operate at a much lower cost compared to GPT-4.5, OpenAI confirmed. As a result, the company will phase out the GPT-4.5 API preview in July, as GPT-4.1 delivers equal or superior performance across key tasks.

Back in February, OpenAI released the GPT-4.5 research preview to a limited group of users and promised broader availability in the weeks to come.

Don't forget to follow Telsat News to stay up to date with technology.