GPT-4.1 is a large language model family released by OpenAI on April 14, 2025. Designed as an evolution of the GPT-4o series, the suite consists of three distinct models: GPT-4.1 (flagship), GPT-4.1 mini, and GPT-4.1 nano. The models are specifically optimized for software engineering tasks, structured instruction following, and long-context processing.
The flagship GPT-4.1 model features a 1 million token context window, representing a significant expansion over previous iterations. It is a multimodal foundation model capable of processing both text and image inputs. Unlike the reasoning-focused o-series, GPT-4.1 is engineered for lower latency and improved cost efficiency in practical developer workflows and agentic applications.
Key performance improvements include a verified score of 54.6% on the SWE-bench, highlighting its proficiency in solving real-world GitHub issues. The model also demonstrates enhanced instruction adherence, particularly for complex formatting requirements and negative constraints. The knowledge cutoff for the GPT-4.1 series is June 2024.