GPT-4o mini, announced today by OpenAI, is available simultaneously in Azure AI, supports text processing functions with excellent speed, and images, audio and video will come later.
We’re also announcing security features by default for GPT-4o mini, extended data residency and service availability, plus performance upgrades to Microsoft Azure OpenAI.
The GPT-4o mini enables customers to deliver stunning applications at a lower cost with lightning speed. The GPT-4o mini is significantly smarter than the GPT-3.5 Turbo – scoring 82% in Measuring Massive Multitask Language Understanding (MMLU) compared to 70% – and is more than 60% cheaper.1 The model provides an expanded 128K pop-up window and integrates the improved multilingual capabilities of GPT-4o, bringing higher quality to languages from around the world.
GPT-4o mini, announced today by OpenAI, is available simultaneously in Azure AI, supports text processing functions with excellent speed, and images, audio and video will come later. Try it for free in the Azure OpenAI Studio Playground.
We are most excited about the new customer experiences that can be enhanced with the GPT-4o mini, especially streaming scenarios such as assistants, code interpreter and search that will benefit from this model’s capabilities. For example, we observed incredible speed when testing the GPT-4o mini on GitHub Copilot, an AI pair programmer that helps you by providing suggestions for code completion in small pauses between keystrokes and quickly updates recommendations with each new character typed.
We’re also announcing updates to the Azure OpenAI Service, including security enhancements by default for GPT-4o mini, expanded data residency and global continuous availability, plus performance upgrades.
Azure AI brings security by default for GPT-4o mini
Security remains paramount to the productive use and trust that we and our customers expect.
We are pleased to confirm that our Azure AI Content Safety features – including Quick Shields and Protected Material Detection – are now “on by default” for use with GPT-4o mini in Azure OpenAI.
We’ve invested in improving the throughput and speed of Azure AI Content Safety features—including the introduction of an asynchronous filter—so you can maximize gains in model speed without compromising security. Azure AI Content Safety already supports developers across industries to protect their generative AI applications, including game development (Unity), tax filing (H&R Block) and education (South Australia Department of Education).
In addition, the GPT-4o mini will be covered by our Customer Copyright Commitment, which gives you the assurance that Microsoft will defend customers against third-party intellectual property claims related to the output content.
Azure AI now offers data residency for all 27 regions
From day one, Azure OpenAI is covered by Azure’s data residency commitments.
Azure AI gives customers both flexibility and control over where their data is stored and processed, offering a complete data residency solution that helps customers meet their unique compliance requirements. We also provide a choice between a hosting structure that meets business, application and compliance requirements. Regional pay-as-you-go and Provisioned Throughput Units (PTUs) offer control over data processing and storage.
We’re excited to announce that Azure OpenAI is now available in 27 regions, including Spain, which launched earlier this month as our ninth region in Europe.
Azure AI announces global pay-as-you-go with highest throughput limits for GPT-4o mini
The GPT-4o mini is now available through our global ongoing deployment at 15 cents per million input tokens and 60 cents per million output tokens, significantly cheaper than previous frontier models.
We’re excited to announce the general availability this month of Global Continuous Deployment, which allows customers to pay for the resources they consume, making it flexible for different workloads while routing traffic globally for higher throughput while still offering control above where the data is at rest.
Additionally, we recognize that one of the issues customers face with new models is the inability to upgrade between model versions in the same region as their existing deployments. Now, with global ongoing deployment, customers will be able to upgrade from existing models to the latest models.
Global pay-as-you-go offers customers the highest possible scale, offering 15 million tokens per minute (TPM) throughput for GPT-4o mini and 30 million TPM throughput for GPT-4o. Azure OpenAI offers GPT-4o mini with 99.99% availability and the same top speed as our partner OpenAI.
Azure AI offers superior performance and flexibility for GPT-4o mini
Azure AI continues to invest in making AI tasks more efficient across Azure OpenAI.
GPT-4o mini is coming to Azure AI with availability in our Batch service this month. Batch provides high-throughput jobs with 24-hour turnaround at a 50% discount rate using off-peak capacity. This is only possible because Microsoft runs on Azure AI, which allows us to make off-peak capacity available to customers.
This month we are also releasing a fine-tuning for the GPT-4o mini that allows customers to further customize the model for your specific use case and scenario to deliver exceptional value and quality at unprecedented speeds. Following our update last month to switch to token-based training billing, we’ve reduced hosting fees by up to 43%. Coupled with our low cost-per-derivation, this makes fine-tuned Azure OpenAI Service deployments the most cost-effective offering for customers with production workloads.
With over 53,000 customers turning to Azure AI to deliver breakthrough experiences at impressive scale, we’re excited to see innovation from companies like Vodafone (customer agent solutions), University of Sydney (AI assistants) and GigXR ( virtual patients with artificial intelligence). ). More than 50% of the Fortune 500 build their applications using Azure OpenAI Service.
We can’t wait to see what our customers do with GPT-4o mini on Azure AI!
1GPT-4o mini: advanced cost-effective intelligence | OpenAI