We are thrilled to announce the addition of OpenAI’s latest models, o1-preview and o1-mini, to the Microsoft Azure OpenAI Service, Azure AI Studio, and GitHub Models. The o1 series is designed to enhance capabilities in complex coding, mathematical reasoning, brainstorming, and comparative analysis, establishing a new standard for AI-powered solutions.
Azure OpenAI Service
Build Custom Generative AI Applications
The o1-preview and o1-mini models are now available in Azure AI Studio and GitHub Models for a select group of Azure customers. This access allows users to collaboratively explore the unique strengths of each model. The o1 series excels in various advanced reasoning tasks, including:
- Complex Code Generation: These models can generate algorithms and perform advanced coding tasks to assist developers.
- Advanced Problem Solving: Ideal for comprehensive brainstorming sessions and tackling multifaceted challenges.
- Complex Document Comparison: Perfect for analyzing contracts, case files, or legal documents to identify subtle differences.
- Instruction Following and Workflow Management: Particularly effective for managing workflows that require shorter context.
Insights from Early Adopters
Early adopters have shared valuable insights regarding the capabilities of these models. For instance, GitHub Copilot, the leading AI developer tool, has been testing the o1 models with promising results in code analysis and optimization.
Thomas Dohmke, CEO of GitHub, stated, “With o1 and its strong reasoning capabilities, GitHub Copilot will enable developers to build for the bigger picture faster. Nothing beats the feeling when you solve a coding problem within minutes instead of hours.
We can’t wait to see what developers do with o1 in their applications.”
Harvey, a leader in generative AI for professional services, is integrating its platform with Azure OpenAI Service. Winston Weinberg, co-founder and CEO of Harvey, remarked, “In our testing, we believe o1 marks a step change in legal reasoning and will enable us to create custom solutions that tackle more ambitious workflows.”
Cognition, an applied AI lab building end-to-end software agents, is also excited about integrating the new models into Devin, the world’s first fully autonomous AI software engineer. Scott Wu, CEO of Cognition, expressed enthusiasm about how innovations like o1 will enhance Devin's ability to solve increasingly complex coding tasks.
Commitment to Responsible AI
Safety remains a top priority as we introduce new generative AI models. The o1 and o1-mini models in Azure OpenAI Service come with “on-by-default” Content Safety features at no additional cost. OpenAI has invested in advanced safety measures that help the model refuse unsafe requests, making the o1 series among the most robust models deployed to date.
We are also expanding our Content Safety systems across all models. New features like Prompt Shields and Protected Materials for Text are now generally available in Azure OpenAI Service. These tools help customers maintain the safety and integrity of their generative AI deployments.
Exploring the New Models
We invite select customers to try out the o1-preview and o1-mini models in the Azure AI Studio playground. This exclusive access allows early experimentation with their advanced reasoning capabilities, with API access expected soon. Interested users can visit the Azure AI Studio model catalog to apply for access today.
![Preview]()
Credit
Model Lineup Overview
The o1 series complements our existing lineup of models:
- o1-preview: Focused on advanced reasoning and solving complex problems across math and science tasks.
- o1-mini: A smaller and faster model that is 80% cheaper than o1-preview, excelling at code generation and small context operations.
- GPT-4o: A versatile multimodal model that excels in text and image processing.
- GPT-4o Mini: A cost-effective version optimized for resource-constrained environments.
- DALL-E: Generates images from text prompts.
- Whisper: Transcribes and translates speech into text.
Conclusion
The introduction of OpenAI's o1-preview and o1-mini models represents a significant advancement in AI capabilities on Microsoft Azure. With their focus on complex reasoning tasks and enhanced safety features, these models are poised to transform how developers approach problem-solving across various domains. We look forward to seeing how our community leverages these innovations to build powerful applications.
For more information on accessing these models or to explore their capabilities further, please visit Azure OpenAI Service.