Fine Tune GPT-4o on Azure OpenAI Service

Get excited! You can now fine-tune GPT-4o using Azure OpenAI services!

We are excited to announce the public preview of fine-tuning for GPT-4o on Azure. After a successful private preview, GPT-4o is now available to all Azure OpenAI customers,Parallelized customization and performance on Azure OpenAI services.

Why Fine Tuning Matters

Fine-tuning is a powerful tool that allows you to tailor advanced models to your specific needs. Whether you want to improve the accuracy of your responses, ensure that your output matches your brand voice, reduce token consumption or latency, or optimize your model for a specific use case, fine-tuning allows you to customize a best-in-class model with your own proprietary data.

GPT-4o: Build better models with your own training data

GPT-4o delivers the same performance as GPT-4 Turbo, but with improved efficiency, and is the best performing OpenAI model on non-English language content. With the release of FineTune for GPT-4o, you can now customize it to your unique needs. FineTune GPT-4o allows developers to train the model on domain-specific data to produce more relevant, accurate, and contextual output.

This release marks a major milestone for Azure OpenAI Service, enabling you to build highly specialized models that drive better performance, achieve higher accuracy using fewer tokens, and create truly differentiated models that support your use cases.

Fine tuning features

Today, we are announcing the availability of text-to-text fine-tuning for GPT-4o, supporting advanced features that help you build custom models to fit your needs, in addition to basic customization.

• Call tool: Include function and tool calls in your training data to allow your custom model to do more!

• Continuous fine-tuning: Update or improve the accuracy of a previously fine-tuned model by fine-tuning it with new or additional data.

• Deployable snapshots: No need to worry about overfitting, now you can distribute the preserved snapshots from each epoch and evaluate them alongside the final model.

• Built-in safety devices: GPT-4, 4o, and 4o mini models have built-in automatic guardrails to prevent the fine-tuned models from generating harmful content.

GPT-4o is available to customers using Azure OpenAI resources in the North Central US and Central Sweden. Stay tuned as we add support in additional regions.

Reduced prices to increase access to experiments

We’ve heard your feedback about model tuning and hosting costs. To make it easier for you to experiment with and deploy your fine-tuned models, we’ve updated our pricing structure:

These changes make experimentation easier and more cost-effective than ever. Updated pricing for the Fine tuning model can be found here. Azure OpenAI Service – Pricing | Microsoft Azure

Get started today!

Whether you’re new to fine-tuning or an experienced developer, getting started with Azure OpenAI Service is easier than ever. Fine-tuning is available through Azure OpenAI Studio and Azure AI Studio, offering a user-friendly interface for those who prefer a GUI, and powerful APIs for advanced users.

Are you ready to get started?