Fine-tuning
Training a pre-built model further on your own data, to make it better at a specific task or voice.
We are working on a detailed page for Fine-tuning - covering why it matters, how it works, related terms, and the tools that use it.
Related terms
From the glossaryFrequently asked questions
Is fine-tuning the same as training from scratch?+
No. Training from scratch builds all model knowledge from random weights using massive datasets and compute. Fine-tuning starts from an already capable model and nudges it toward a specific style or domain.
How much data do I need to fine-tune?+
Far less than pre-training. Hundreds to a few thousand high-quality examples are often enough to shift style or add domain vocabulary. More data helps for complex tasks.
When should I use fine-tuning instead of prompting?+
When the desired behaviour is consistent and hard to describe in a prompt alone, when you need a specific tone or format every time, or when latency and cost make long system prompts impractical.