Introducing GPT-4o-2024-08-06 API with Structured Outputs on Azure

We are excited to announce the release of the API for OpenAI’s latest and most advanced AI model, GPT-4o-2024-08-06. Azure OpenAI Service. This model brings significant advances in AI capabilities, including groundbreaking JSON structured output capabilities that enhance developer productivity and application diversity. Starting today, you can leverage the full capabilities of GPT-4o-2024-08-06 in your generative AI applications.

Three Key Highlights of GPT-4o-2024-08-06

GPT-4o-2024-08-06 is a groundbreaking advancement in GenAI model technology, boasting several notable new features and improvements.

1. New feature: JSON structured output

JSON Structured Output allows developers to specify the desired output format from their AI models by defining a JSON schema. This feature simplifies the creation of structured data and reduces the need for post-processing by ensuring that the output is well-structured and consistent. It also provides significant cost savings, with input costs reduced by up to 50% and output costs reduced by up to 33% compared to its predecessor, GPT-4o-2024-05-13.

2. Improved cost efficiency

The GPT-4o-2024-08-06 model provides significant cost-saving benefits to developers.

• Input Cost: Up to 50% reduction ($2.50 per 1M input tokens)
• Output Cost: Up to 33% reduction ($10.00 per 1M output tokens)

3. Extended availability

The GPT-4o-2024-08-06 API is now available globally with global and regional standards deployments in all US regions and Central Sweden.

Why choose GPT-4o-2024-08-06?

The GPT-4o-2024-08-06 model is designed to perform a wide range of tasks with minimal cost and latency, making it suitable for applications that require fast, real-time text responses. The introduction of JSON structured output provides 100% confidence in the evaluation, ensuring that the output matches the defined output schema perfectly. This innovation improves the efficiency and accuracy of AI-based applications in a variety of use cases.

Structured Output: Revolutionizing AI Development

What is structured output? Structured output allows developers to define the exact output format using JSON schema, making it easier to generate data with the desired structure for a variety of applications. This feature is ideal for seamless integration with other systems.

Key benefits for developers:

• Consistent data format: The output adheres to a defined JSON schema, minimizing the need for manual post-processing.
• Flexible tool integration: Supports specific function signatures to provide precise output for tool integration.
• Simplified application development: Simplify data payload creation to improve overall user experience.

Example use cases: Customer support chatbots can use structured output to generate responses specifically formatted for logging and analytics. This approach enables seamless data integration, streamlining the process of logging, analytics, and taking action based on data.

Getting started with structured output: To take advantage of structured output, follow these three simple steps:

1. JSON Schema Definition: Determines the structure of the AI ​​output.
2. Build AI models: Leverage structured output features in API calls.
3. Integration and Testing: Integrate the results into your application and test them thoroughly.

Availability and Pricing

Pricing Options: The GPT-4o-2024-08-06 API offers significant cost savings over its predecessor, GPT-4o-2024-05-13. The input is $2.50 per 1M tokens and the output is $10.00 per 1M tokens. For detailed pricing, please refer to: Azure OpenAI Services Pricing Page.

Availability by region:

We are excited to offer APIs in all US regions and Central Sweden as part of our global and regional standards deployment.

Comparing GPT-4o-2024-08-06 to other models

Compared to GPT-4o-2024-05-13: The new model improves performance and robustness by providing advanced features such as structured output for more accurate and structured output in a variety of applications.

Compared to GPT-4o Mini: While both models share multimodal capabilities, GPT-4o Mini is optimized for cost-efficiency, making it an ideal choice for tests and applications where cost is a critical factor.

Next steps for customers

Learn more Models in our documentation And try the model Azure AI Studio. If you are currently using GPT-4o-2024-05-13, you can easily upgrade to GPT-4o-2024-08-06 if both models are available in your region and offering type. There are currently no plans to automatically upgrade to the new model, but we will notify you if there are any changes in the future.