Steve Sweetman and Naomi Moneypenny announce the public preview of OpenAI’s Sora 2 in Azure AI Foundry, explaining its advanced video generation features and secure integration options for developers and creators.

Sora 2 Public Preview Now Available in Azure AI Foundry

Authors: Steve Sweetman and Naomi Moneypenny

OpenAI’s Sora 2, a powerful multimodal video generation model, is now available in public preview through Azure AI Foundry. This makes it easier than ever for developers and organizations of all sizes to build, prototype, and deliver innovative video-based applications directly from Microsoft’s unified cloud AI platform.

Key Features

Realistic Video Generation: Advanced world simulation and physics underpin highly realistic output capabilities.
Multimodal Inputs: Generate videos using text, images, or even other videos as prompts.
Synchronized Audio: Automatically generate and synchronize dialogue and audio tracks, with support for multiple languages.
Enhanced Creative Control: Fine-tune camera angles, scene details, and prompt settings for precision outputs.
Security and Safety: Built-in input and output content filtering, plus Microsoft Azure’s enterprise-grade security, compliance, and governance features.
Integration: Seamless integration with business workflows through API access.

Use Cases

Sora 2 in Azure AI Foundry opens up possibilities for a variety of roles:

Marketers: Rapidly create high-quality video assets for campaigns.
Retailers: Design interactive, personalized content for customer engagement.
Creative Directors: Prototype dynamic scenes and cinematic experiences, leveraging world simulation for concept development.
Educators: Author immersive instructional media and interactive lesson plans for better learner engagement.

Access, Pricing, and Model Catalog

Now in Public Preview: Sora 2 is available via API through Standard Global in Azure AI Foundry.
Pricing: .10 per second for video generation (720×1280 portrait, 1280×720 landscape).
Wide Model Catalog: Azure AI Foundry also features models like GPT-image-1, Black Forest Lab’s Flux 1.1, Kontext Pro, and more for additional generative capabilities.

For more technical documentation and updates, see the Azure AI Foundry Models page.

Security, Compliance, and Responsible AI

Microsoft prioritizes responsible and safe AI deployment. Sora 2 benefits from:

Input Filtering: Prompts (text, image, video) are screened for compliance and safe content.
Output Filtering: Video and audio results are analyzed to block restricted content.
Privacy Controls: Organizational and customer data is protected within Azure’s privacy and governance frameworks.

More on Microsoft’s responsible AI principles can be found here.

Getting Started

To begin working with Sora 2 in Azure AI Foundry:

Visit the Azure AI Foundry catalog
Review technical requirements, pricing, and deployment types
Access models via API to integrate video generation into applications

Sora 2 aims to accelerate the creative process for developers, allowing rapid prototyping and ideation for compelling visual storytelling, all secured and managed through Microsoft’s enterprise-ready platform.

This post appeared first on “The Azure Blog”. Read the entire article here