gpt-4.1 · openai

GPT-4.1 is a flagship large language model optimized for advanced instruction following, practical software engineering and long-context reasoning. It features a 1 million-token context window and outperforms GPT-4o and GPT-4.5 in benchmarks including programming (54.6% SWE-bench Verified), instruction following (87.4% IFEval) and multimodal comprehension. Fine-tuned for accurate code diff processing, agent reliability and high recall in lengthy document contexts, it serves as an ideal option for agents, IDE tools and enterprise knowledge retrieval.