# OpenAI o-series Reinforcement Fine-Tuning (RFT): Grader Function Design

> Source: https://sukruyusufkaya.com/en/learn/fine-tuning-cookbook/ftc-openai-rft-grader-function
> Updated: 2026-05-14T14:43:00.081Z
> Category: Fine-Tuning Cookbook (Model-by-Model)
> Module: Part XIV — Closed-Source API Fine-Tuning
**TLDR:** OpenAI announced RFT in late 2024: fine-tune o-series models (o1, o3, o4-mini) with reasoning RL. **Grader function** — function you write that gives numerical score to model output (math correctness, code execution, custom rule). Ideal for verifiable domains. JSON-based grader spec.