OpenAI Releases the Reinforcement Fine – Tuning Function.

OpenAI has introduced Reinforcement Fine-Tuning (RFT) to enhance model performance in complex domains, particularly in areas such as coding, scientific research, and finance. RFT combines chain-of-thought reasoning with task-specific scoring, enabling organizations to customize expert-level models. Additionally, users can now perform supervised fine-tuning on GPT-4.1 nano to better meet specific requirements. RFT is currently available to verified organizations, and users who share datasets may receive discounts while contributing to model improvement. Reinforcement fine-tuning has also shown strong potential in fields like medical diagnosis, where it can improve accuracy by analyzing symptoms and test results.