Note
Access to this page requires authorization. You can try signing in or .
Access to this page requires authorization. You can try .
What's new in Microsoft Foundry?
Welcome! This article highlights key article and product updates in Microsoft Foundry for May 2026.
New articles
- Azure OpenAI
- Quickstart: Create a provisioned throughput deployment
- Determine provisioned throughput unit (PTU) sizing for a workload
- Auto and direct model routing with the Responses API
- Use model router with Foundry agents
- Automate Azure OpenAI deployments with quota
- Get started with Azure OpenAI audio generation
- Foundry Agent Service
- Enable incoming A2A on a Foundry agent (preview)
- Routines in Microsoft Foundry (preview)
- Automate agents with routines (preview)
- Build a voice agent with hosted agents (preview)
- Add managed MCP servers powered by connector namespaces (preview)
- Connect agents to Microsoft Fabric with Fabric IQ (preview)
- Connect agents to Microsoft 365 with Work IQ (preview)
- Enable tool search in a toolbox (preview)
- What is the agent optimizer? (preview)
- Quickstart: Optimize a hosted agent (preview)
- Make your agent optimizer-ready (preview)
- Create an evaluation dataset for the agent optimizer (preview)
- Optimize agent instructions, skills, tools, and models (preview)
- Evaluations and observability
- Rubric evaluators (preview)
- Run agent evaluations with the azd CLI (preview)
- Run benchmark evaluations in Microsoft Foundry (preview)
- Generate a synthetic evaluation dataset (preview)
- Review agent interactions with Trace Replay (preview)
- Convert agent traces into evaluation datasets (preview)
- Microsoft Foundry Tracing and Data Handling
- Guardrails
- Foundry Models and platform
Updated articles
- Azure OpenAI
- What is provisioned throughput for Foundry Models?
- Provisioned throughput billing and cost management
- Operate provisioned deployments in production
- Manage traffic with spillover for provisioned deployments
- Latency and performance optimization
- Monitor Azure OpenAI reference
- Priority processing for model deployments
- Quota management and rate limits
- Model router for Microsoft Foundry concepts
- How to use model router for Microsoft Foundry
- Foundry Agent Service
- Connect to an A2A agent endpoint from Foundry Agent Service
- Deploy a hosted agent
- Hosted agents in Foundry Agent Service (preview)
- Tool support by model and region
- Use skills with Microsoft Foundry agents (preview)
- Automate browser tasks with Foundry agents
- Curate intent-based toolbox in Foundry (preview)
- What is Microsoft Foundry Agent Service?
- Agent development lifecycle
- Quickstart: Deploy your first hosted agent
- Evaluations and observability
- Foundry Models and platform
Feedback
Was this page helpful?