We are excited to announce that o1 models will soon be available on the Microsoft Azure OpenAI service. This multimodal model provides advanced inference capabilities and enhancements that significantly enhance AI applications and solutions. o1 models support both text and vision input, making them ideal for a wide range of applications, from complex problem solving to contextual AI experiences.
We are excited to announce that o1 models will soon be available on the Microsoft Azure OpenAI service. This multimodal model provides advanced inference capabilities and enhancements that significantly enhance AI applications and solutions. o1 models support both text and vision input, making them ideal for a wide range of applications, from complex problem solving to contextual AI experiences.
Customers using o1 models in Azure OpenAI service
Rohirrim: o1 model analysis is “amazing”
Rohirrim Inc. is a fast-growing AI startup that has built Rohan, a secure, knowledge-based generative AI platform designed for Fortune 1000 companies and the aerospace or defense sectors. Rohan simplifies complex RFP responses, eliminating the need to sift through unstructured data and providing fast, accurate answers in your company’s unique voice. Hosted as a dedicated Platform as a Service (PaaS), it ensures data control, high compliance, and low governance risk to improve productivity and reduce burnout.
With o1, we’ve unlocked a layer of contextual understanding and nuanced analysis that, frankly, is truly amazing. This evolution means we can now apply AI to the complex and detailed aspects of proposal management, while freeing up human experts to focus on the higher-order, intuitive thinking, such as insight and creativity, that truly helps organizations grow. there is. —Steven Aberle, Rohirrim CEO
Wrtn Technologies: Response accuracy has “increased exponentially”
Wrtn Technologies is an AI-based solutions provider recognized for its innovative approach and commitment to excellence. Leveraging the advanced features of the o1 model, Wrtn Technologies transformed its digital presence and performance.
With the o1 model, our measurements show that the success rate of correct answers to queries increases exponentially—Dongjae “DJ” Lee, Chief Product Officer, Wrtn Technologies
Harvey: o1 is a “sophisticated reasoner.”
Harvey is the generative AI platform trusted by law firms and professional service providers. It helps lawyers and other professionals by automating complex legal tasks, such as summarizing and comparing documents, performing due diligence, and consulting case law. Harvey leverages Azure AI infrastructure to improve efficiency and productivity in the legal industry.
(W) What you have is an incredibly powerful model and sophisticated reasoner. You can plan from scratch to solve complex tasks. And there is a degree of control you can control by following instructions. … (You) can insert a human into the loop and actually steer the model toward higher quality output.—Niko Grupen, Director of Applied Research, Harvey
See o1 in action here as Harvey uses it to analyze a series of complex legal documents.
New o1 model features for developers
The o1 model introduces several new features that can improve your inference-based development experience. Some of them have been seen before in other models, while others are new to the o1. These include:
- vision input: With the addition of vision input capabilities, o1 expands its inference capabilities to process and analyze visual data to extract valuable insights and produce comprehensive text output.
- developer message: Developer messages are similar to system messages in the GPT model, so you can use the ‘role’: ‘developer’ attribute to convey instructions and context for a specific use case.
- Inference effort parameters: The inference effort parameter allows you to adjust the cognitive load of your model with low, medium, and high inference level options, giving you more control over model performance.
- Extended Context Window: The o1 model now boasts an extended context window of 200K tokens and a maximum output of 100K tokens, providing ample room for complex and detailed responses.
- structured output:o1 now adds support for structured output, which helps simplify the creation of well-defined, structured output limited to a JSON schema.
- equipment: We can now use the o1 model output to generate input to the function, just like the previous model.
- low latency: o1 uses on average 60% fewer inference tokens than o1-preview for a given request.
With the introduction of tool support, these models can power agent AI applications, allowing them to plan, reason, and implement agent tools that automate business processes. The newly added o1 capabilities can deliver significant benefits to developers by increasing customization, control, capacity, and insight into AI-based scenario inference. These advancements in the Azure OpenAI service allow developers to leverage cutting-edge technologies while maintaining enterprise-grade security and global compliance.
Agility and global accessibility with privacy and security
Microsoft is committed to providing customers with the latest OpenAI models through same-day access. The o1 model is no exception and provides the most advanced inference capabilities available today on Azure. Our services are designed to provide broad global reach, one of the best local data choices, and help keep your data private and secure.
- Global reach and local data: Azure OpenAI service provides 99.9% reliability with access to 28 regions to ensure data residency and compliance with local regulations.
- latest innovations: We provide same-day access to the latest OpenAI models, ensuring you stay ahead.
- Enterprise-grade security: Protect your data with built-in security features, including private networking, managed identities, and Azure monitoring.
- Developer-friendly integration: Seamless integration with GitHub, Visual Studio, and Copilot Studio allows developers to easily build and deploy AI solutions.
- Responsible AI: Our services have built-in responsible AI features, including content filters, prompt shields, and grounding detection, to support safe and ethical use of AI.
- What’s new in o1: The o1 model introduces developer messages, inference effort parameters, and extended response objects to increase functionality and customization for developers.
The Azure OpenAI service also introduces several new fine-tuning features, including support for o1-mini models with augmented fine-tuning. This allows organizations to improve performance and reduce costs by customizing AI models to their specific needs. The service also introduces Direct Preference Optimization (DPO), which adjusts model weights based on human preferences, and Store Completion, which captures and stores input-output pairs for fine-tuning. These advancements demonstrate our commitment to delivering powerful, flexible, and efficient AI solutions through Azure OpenAI services.
In addition to these updates, we recently introduced a suite of developer-focused features designed to enhance your AI development experience in your preferred environments like GitHub and Visual Studio Code. Built-in safety tools like Azure AI Foundry, GitHub Models, the new Azure AI Agent service, and evaluation and monitoring allow developers to integrate advanced AI capabilities directly into their applications. These tools provide powerful support for building, testing, deploying, and managing AI models, ensuring the flexibility and control you need to innovate and deliver superior performance with trustworthy AI systems. These new capabilities can help you streamline your workflow, improve efficiency, and open up new possibilities for AI-driven development.
Please join us on this journey
Explore the capabilities of the o1 model and see how it can transform your business. The Azure OpenAI service gives you access to the latest innovations, enterprise-grade security, and global services that ensure your data is private and secure. Join us on this journey and discover the potential of AI to drive your company’s success.
To access the latest models like o1, sign up for Azure AI Foundry today.
