Model Distillation in the API
OpenAI just shipped model distillation on the API—let your smaller models learn from GPT-4 without the $20/MTok bill.

Why it matters
OpenAI is making it operationally cheaper to deploy AI at scale by letting developers distill frontier model knowledge into smaller, cheaper models directly within the platform. This reshapes the cost-performance tradeoff for production AI applications.
The key facts
5 to knowOpenAI launches model distillation feature on API platform
Allows fine-tuning of cost-efficient models using frontier model outputs
Reduces inference costs for production deployments
Distillation workflow available natively on OpenAI platform
Published October 1, 2024
Go to the source
OpenAI Blogopenai.com
Publisher excerpt: Fine-tune a cost-efficient model with the outputs of a large frontier model–all on the OpenAI platform
