Microsoft Foundry (formerly Azure AI Foundry)
Microsoft's Azure hub for enterprise models and agents
Overview
Microsoft's enterprise AI platform (renamed from Azure AI Foundry to Microsoft Foundry in January 2026): a model catalog spanning OpenAI, Anthropic, Meta and Mistral, Foundry Agent Service for building agents, and evaluation tied into Azure governance. For enterprise Azure developers.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 13 Jul 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
What you can produce with Microsoft Foundry (formerly Azure AI Foundry)
- Model catalog spanning OpenAI, Anthropic, Meta, and Mistral
- Foundry Agent Service for building and orchestrating agents
- Evaluation and observability tooling
- Prompt flow and orchestration for multi-step apps
- Azure RBAC, governance, and content safety integration
- GitHub and VS Code developer tooling
- Pay-per-token billing on a free platform layer
ASEAN Perspective
Microsoft Foundry (formerly Azure AI Foundry) in Southeast Asia
ASEAN-region availability and pricing notes coming soon. Drop the editorial team a note via /contact/ if you can supply local context (Singapore/Malaysia/Indonesia/Thailand/Vietnam).
Renamed from Azure AI Foundry to Microsoft Foundry in January 2026 — the name changed, the product and your deployments didn't. It's Microsoft's enterprise AI hub: a model catalog that now spans OpenAI, Anthropic's Claude, Meta, and Mistral, the Foundry Agent Service for building and orchestrating agents, plus evaluation and observability tooling, all bolted to Azure's governance and RBAC stack. The platform itself is free to use; you pay per-token for the models and features you actually consume. Reviewers rate it well but are candid that setup is intricate, the learning curve is steep, and costs climb for anything beyond a pilot — plus operational chores like managing tokens-per-minute quotas and multi-region deployment take real effort. Best fit for larger enterprises already committed to Azure with the technical depth to run it. Teams that want model choice without leaving Microsoft's cloud will feel most at home; startups on a tight budget less so.
What people say
Microsoft Foundry holds a 4.6-star rating across 124 reviews on G2 — strong for an enterprise platform — and the January 2026 rename from Azure AI Foundry was cosmetic: existing deployments carried over automatically and functionality stayed put.
What reviewers value most is the agent tooling and the model catalog. They describe a development environment that pulls models, tools, and frameworks into one place, and they call out the orchestration features for coordinating multiple AI agents in a single workflow. The catalog's freshness gets specific praise — new models land quickly and the APIs behave as documented — which matters to teams that want access to OpenAI and Anthropic models without stitching together separate vendor integrations. Scalability and customization for specific business needs round out the positives.
The criticism is consistent and material. Setup is described as intricate and time-consuming, with a steep learning curve that assumes real AI-development expertise; this is not a platform reviewers recommend to non-technical teams. Cost is the other recurring theme — several reviewers flag it as a barrier for startups or budget-constrained projects, since the token-based billing (GPT-class models around $2.50 per million input and $10 per million output tokens, with rates varying widely across the catalog) adds up under production load. Operational friction shows up too: users report effort around deploying to multiple regions, reducing latency, managing tokens-per-minute limits, and security configuration.
The consensus positioning: Microsoft Foundry is aimed squarely at larger enterprises and technically capable developers building integrated AI systems, and for organizations already standardized on Azure it's a natural fit that combines model choice with Microsoft's governance and compliance tooling. The pattern in reviews is that the platform's capability is rarely the problem — it's the configuration overhead and the cost trajectory that teams underestimate. Prototype on the free platform layer, model the per-token and quota costs against your real workload, and budget for a ramp-up period before treating it as production infrastructure.
Summary of public user & expert reviews, compiled by RECATOOLS.
About this listing
This entry was compiled from publicly available data including Microsoft Foundry (formerly Azure AI Foundry)'s official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with Microsoft Foundry (formerly Azure AI Foundry) unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to Microsoft Foundry (formerly Azure AI Foundry) directly →
Spotted something out of date? Suggest an update →
Alternatives to Microsoft Foundry (formerly Azure AI Foundry)
More in Code & Dev Tools
