> For the complete documentation index, see [llms.txt](https://docs.aisuru.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aisuru.com/en/advanced-features/fine-tuning/what-is-fine-tuning-and-what-is-it-for.md).

# What is fine-tuning and what is it for

Fine-tuning is a machine learning process that involves taking a pre-trained model and adapting it to a specific task or particular domain. This advanced technique is based on the idea of leveraging the knowledge a pre-trained model has already acquired and adapting it to a specific purpose.

### What are the benefits of fine-tuning?

1. **Improves performance**: fine-tuning lets you sharpen the model's capabilities on specific tasks, leading to more accurate and relevant results.
2. **Lets you scale costs**:
   1. Fine-tuning is cheaper than training a model from scratch;
   2. Fine-tuning is generally performed on smaller, more cost-effective language models (LLMs): this means that for equivalent responses, each message will cost you significantly less.

Fine-tuning also requires relatively little data: it's possible to achieve good results with just a few thousand examples.

### The fine-tuning process

1. **Choosing the model** -> generally, a small, general-purpose model is chosen (like gpt-4o mini);
2. **Preparing the data** -> a specific dataset of examples is collected to feed to the language model (LLM); [with AIsuru, preparing data for fine-tuning is extremely straightforward](/en/advanced-features/fine-tuning/how-to-do-fine-tuning-with-aisuru.md);
3. **The actual fine-tuning**: on the Microsoft Azure or [OpenAI](https://platform.openai.com/finetune) platform, you'll fine-tune the model you've chosen.

✅ Done! Once you've completed these steps, you can immediately start using the [fine-tuned model in your AIsuru Agent](/en/advanced-features/fine-tuning/using-fine-tuned-models-on-aisuru.md)!

### When to use fine-tuning

Fine-tuning is particularly useful when you want to modify the behavior and increase the accuracy of a specific model.

By using fine-tuning, **you can get models that are more precise, relevant, and suited to the specific needs of a project or organization**, while still leveraging the power of models from major players like OpenAI.