Microsoft is pushing the boundaries of artificial intelligence with significant enhancements to its Azure AI Foundry platform, designed to empower organizations and transform industries. These latest updates, including the introduction of powerful new models like OpenAI’s GPT-4.5, advanced fine-tuning techniques, and new enterprise tools for agents, are designed to accelerate the journey from AI experimentation to delivering tangible business outcomes.
Introducing GPT-4.5 on Azure OpenAI Service
Building on the success of previous models, GPT-4.5 is the newest and most capable general-purpose model available. This research preview showcases the achievements from scaling pre and post-training, a significant advancement in unsupervised learning techniques.
GPT-4.5 offers a more natural interaction experience. It possesses a broader knowledge base and features improved ‘EQ,’ making it adept at coding, writing, and problem-solving tasks. Compared to GPT-4o, GPT-4.5 demonstrates a lower hallucination rate of 37.1% versus 61.8% and higher accuracy of 62.5% compared to 38.2%, providing developers with more precise and relevant results. Enhanced alignment techniques further improve its ability to follow instructions, understand nuances, and engage in natural conversations.
Developers can leverage GPT-4.5 in many ways to boost productivity and creativity. Some applications include:
- Clear and effective communication: crafting emails, messages, and documentation.
- Personalized learning: Offering learning and coaching experiences to acquire new skills.
- Brainstorming: Generating new ideas and solutions.
- Project planning and execution: Organizing tasks and ensuring thoroughness.
- Automated tasks: Simplifying processes and workflows.
- Coding: Providing step-by-step guidance and automating repetitive tasks.
Starting today, enterprise customers can utilize GPT-4.5 in Azure AI Foundry. It is also available to GitHub Copilot Chat to Copilot Enterprise users.
New Models: Phi-4, Stability AI, and Recent Releases
The latest AI models are focused on delivering specialized capabilities with greater efficiency, representing a shift towards purpose-built AI. Here are some notable launches on Azure AI Foundry:
Microsoft’s Phi models continue to push the boundaries of what’s possible with smaller, more efficient architectures:
- Phi-4-multimodal unifies text, speech, and vision for context-aware interactions. Retail kiosks can diagnose product issues via camera and voice inputs.
- Phi-4-mini packs impressive performance into just 3.8 billion parameters with a 128K-token context window. It outperforms larger models on coding and math tasks while increasing inference speed by 30% compared to previous models.
Stability AI continues to advance generative imaging with models that accelerate creative workflows:
- Stable Diffusion 3.5 Large generates high-fidelity marketing assets faster than previous versions, thus maintaining brand consistency across diverse visual styles.
- Stable Image Ultra achieves photorealism for product imagery, reducing photoshoot costs through accurate material rendering and color fidelity.
- Stable Image Core an enhanced version of SDXL, provides high-quality output with exceptional speed and efficiency.
Cohere enhances information retrieval capabilities with its latest ranking technology:
- Cohere ReRank v3.5 delivers more accurate search results through its 4,096-token context window and support for over 100 languages, surfacing relevant content even without exact keyword matches.
The GPT-4o family expands with two specialized variants:
- GPT-4o-Audio-Preview handles audio prompts and generates spoken responses with appropriate emotion and emphasis, which is ideal for digital assistants and customer service.
- GPT-4o-Realtime-Preview which eliminates conversation lag with breakthrough latency reduction, creating human-like interaction flows.
Agora, a pioneer in enabling real-time engagement, has been using GPT-4o-Realtime. Patrick Ferriter, Senior Vice President, Product & Marketing, at Agora, stated, “GPT-4o-Realtime has revolutionized voice interaction for our conversational AI product, empowering developers with multilingual human-like voices, stable streaming, and ultra-low latency across customer service, telemedicine, and education.”
These advances collectively signal AI’s evolution toward more natural, responsive, and efficient interactions across diverse use cases and deployment environments.
New Customization Tools
With the model library exceeding 1,800 offerings, Microsoft is committed to pushing the boundaries of experimentation and observability. The new suite of fine-tuning tools supports the rise of unsupervised learning techniques.
- Distillation workflows: Azure OpenAI Service now offers a code-first approach to model distillation with Stored Completions API and SDK. This strategy allows smaller models to inherit knowledge from larger models like GPT-4.5.
- Reinforcement fine-tuning: A technique, now in private preview, teaches models to reason in new ways by rewarding correct logical paths and penalizing incorrect reasoning.
- Provisioned Deployment for fine-tuning: Azure OpenAI Service offers Provisioned Deployments for fine-tuned models, ensuring predictable performance and costs through Provisioned Throughput Units (PTUs).
- Fine-tuning for Mistral Models: Exclusive to Azure AI Foundry, Mistral Large 2411 and Mistral 3B now support fine-tuning for industry-specific tasks like healthcare document redaction.
Secure Automation and Scale for Enterprise Agents
Security and scalability are strategic imperatives in today’s enterprise landscape. Microsoft is introducing two powerful features:
- Bring your Vnet: Azure AI Agent Service now allows all AI agent interactions, data processing, and API calls to remain securely within your organization’s own virtual network, eliminating exposure to the public internet. Early adopters like Fujitsu are leveraging this to improve sales by 67% with their sales proposal creation agent.
- Magma (Multi-Agent Goal Management Architecture): Available through Azure AI Foundry Labs, Magma revolutionizes complex workflow orchestration by coordinating hundreds of AI agents in parallel. Magma enables tackling large-scale challenges like supply chain optimization with unprecedented speed and accuracy, which bridges the physical and digital agentic world.
Microsoft is excited about these innovations and is looking forward to seeing how developers will harness these technologies to spur innovation.
