[ ABORT TO HUD ]
SEQ. 1
SEQ. 2
SEQ. 3

Fine-Tuning in Foundry

🎯 Fine-Tuning & Customization 10 min 80 BASE XP

The Fine-Tuning Process

Supported Models for Fine-Tuning

ModelMin Training ExamplesTypical Use
o3-mini10Reasoning-focused customization (New in 2026)
GPT-4o / OpenAI o3-mini10High-quality custom behavior
GPT-4o mini10Cost-effective custom models

Global Training (2026 Feature)

As of April 2026, Foundry supports Global Training for models like o3-mini. This allows you to launch fine-tuning jobs across 13+ Azure regions, offering lower per-token training rates compared to standard regional training.

Reinforcement Fine-Tuning (RFT)

For reasoning models (o-series), Foundry provides Reinforcement Fine-Tuning (RFT). Unlike Supervised Fine-Tuning (which teaches formatting or style), RFT aligns model behavior with complex business logic by explicitly rewarding accurate reasoning paths.

Training Data Format (SFT JSONL)

{"messages": [
  {"role": "system", "content": "You are a legal contract analyzer."},
  {"role": "user", "content": "Analyze this NDA clause: ..."},
  {"role": "assistant", "content": "Risk Level: Medium. Key concerns: ..."}
]}

Fine-Tuning Costs

  • Training - Charged per token processed during training
  • Hosting - Hourly fee while the model is deployed (even when idle)
  • Inference - Per-token, typically higher than base models
🎯 Pro Tip: Start with 50-100 high-quality examples for your first fine-tuning run. Quality of examples matters far more than quantity. One perfect example teaches more than 100 mediocre ones.
FOUNDRY VERIFICATION
QUERY 1 // 2
What is Reinforcement Fine-Tuning (RFT) designed to do?
Change the model's language
Align model behavior with complex business logic by rewarding accurate reasoning
Reduce the model's memory footprint
Generate images