Microsoft Foundry — Models, Agents and Azure OpenAI
How Microsoft Foundry really works: the Foundry resource and its projects (and what hub-based projects still do), the October 2026 models (GPT-6.1, GPT-6, GPT-5.6, Claude 5.5), deployment types from Global Standard to Data Zone, PTU and Batch, TPM and RPM quotas and the four kinds of 429, spillover, guardrails and Prompt Shields, evaluations, Foundry Agent Service, private endpoints, and the model retirement calendar.
Sign up free to read the full topic
More Cloud topics
- Managed Service for Apache Spark (formerly Dataproc)
- Managed Service for Prometheus
- MemoryDB — Durable In-Memory Database
- Microsoft Defender for Cloud — CSPM and Workload Protection
- Microsoft Entra ID — Identity and Access
- Microsoft Fabric
- Microsoft Sentinel — Cloud-Native SIEM and SOAR
- Migrate to Virtual Machines and Migration Center
- Migration Hub — Migration Planning and Tracking
- MIGs — Compute Engine managed instance groups
- Multi-Cloud & Hybrid Architecture
- Neptune — Managed Graph Database
Start free on Cariara · Coding problems · System design studio