In a significant leap forward for AI development, OpenAI has unveiled its newest suite of language models: GPT-4.1, along with compact versions GPT-4.1 Mini and GPT-4.1 Nano. These new models boast powerful improvements in coding performance, instruction-following accuracy, and long-context comprehension, positioning them ahead of the company’s previous flagship, GPT-4o.
Available exclusively through OpenAI’s API, the GPT-4.1 series delivers robust performance while offering enhanced real-world utility — a priority, according to OpenAI CEO Sam Altman.
What’s New in GPT-4.1?
- Coding Capabilities: GPT-4.1 outperforms previous models with a 21% improvement over GPT-4o and 27% over GPT-4.5 in coding benchmarks.
- Long Context Understanding: The model supports up to 1 million tokens, enabling deeper, more complex document and data comprehension
- Instruction Following: Enhanced ability to follow nuanced instructions makes GPT-4.1 ideal for powering autonomous AI agents and tools.
- Lower Cost: Despite the upgrades, GPT-4.1 operates at a significantly reduced cost compared to GPT-4.5, making it more accessible to developers and businesses.
- Updated Knowledge Base: GPT-4.1 comes equipped with refreshed knowledge up to June 2024, keeping its insights current and relevant.
N also confirmed that it will retire the GPT-4.5 preview available via API in July 2025, as GPT-4.1 meets or surpasses its predecessor in performance.
This launch comes just months after OpenAI offered a limited preview of GPT-4.5 in February, aimed at select developers. The transition to GPT-4.1 marks a strategic pivot toward more powerful, efficient, and developer-friendly AI tools.
