OpenAI announced the expansion of its reinforcement fine-tuning research program, part of its “12 Days of OpenAI” initiatives in December.

This program is designed to assist developers and researchers in creating expert models for complex, domain-specific tasks.

The initiative allows developers to use a technique to customize models by applying high-quality tasks and reference answers.

Applications are open to research institutions, universities, and enterprises from various fields.

OpenAI plans to make this capability available to the public by early 2025.

Recent OpenAI developments