# Microsoft Foundry (Azure AI Foundry) — AI Compare Ninja

> Build, ground, and manage AI models and agents on Azure; each service is billed separately.

## Overview
- **Category:** Development & Code (Agent & RAG frameworks)
- **Status:** Active · checked - Active platform named Microsoft Foundry. Azure AI Foundry is its previous name.
- **Paid from:** 10 / 50 USD / per 1 million input/output tokens (GPT-6 Astra Standard Global (Short Context))
- **Free access:** Free trial
- **Website:** https://azure.microsoft.com/en-us/products/ai-foundry
- **Canonical URL:** https://www.aicompare.ninja/en/tool/azure-ai-foundry

## What it does
- Models, Agent Service, and your data sources in one Azure platform
- Governance, access controls, and evaluation; compliance still depends on configuration
- Standard consumption or separately booked Provisioned Throughput capacity
- Astra example prices distinguish Global and US Data Zone, and short and long context

## In brief
Models, Agent Service, and your data sources in one Azure platform. Microsoft Foundry was previously Azure AI Foundry. There is no single platform price: model, context, deployment, and additional services determine cost. Astra examples are USD list prices, not the cheapest entry. Provisioned Throughput is an alternative capacity booking, not a mandatory surcharge on Standard calls. Search, storage, and networking can cost extra.

For Astra, more than 272,000 input tokens triggers long-context rates for the entire request. Cache reads and writes have separate rates. Quotas and model availability depend on deployment and region. Azure's USD 200 credit for 30 days is for new customers only and excludes Marketplace offers; it is not a permanent free Foundry plan.

Privacy commitments for models sold by Azure do not automatically cover every third-party offer. Global can process requests worldwide; Data Zone confines processing to its zone. No training without permission or instruction does not mean stateful services or abuse monitoring store no data.


## Limits
For Astra, more than 272,000 input tokens triggers long-context rates for the entire request. Cache reads and writes have separate rates. Quotas and model availability depend on deployment and region. Azure's USD 200 credit for 30 days is for new customers only and excludes Marketplace offers; it is not a permanent free Foundry plan.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Azure trial credit | free | 30 days | USD 200 Azure trial credit for new customers, valid 30 days; Excludes Marketplace offers; not a permanent free Foundry plan |
| GPT-6 Astra Standard Global (Short Context) | 10 / 50 USD | per 1 million input/output tokens | Up to 272,000 input tokens; Cache read / cache write: 1 / 12.5 USD; Astra example prices distinguish Global and US Data Zone, and short and long context |
| GPT-6 Astra Standard Global (Long Context) | 20 / 75 USD | per 1 million input/output tokens | Over 272,000 input tokens: rate applies to the entire request; Cache read / cache write: 2 / 25 USD; Astra example prices distinguish Global and US Data Zone, and short and long context |
| GPT-6 Astra Standard Data Zone (US) (Short Context) | 11 / 55 USD | per 1 million input/output tokens | Up to 272,000 input tokens; Cache read / cache write: 1.1 / 13.75 USD; Astra example prices distinguish Global and US Data Zone, and short and long context |
| GPT-6 Astra Standard Data Zone (US) (Long Context) | 22 / 82.5 USD | per 1 million input/output tokens | Over 272,000 input tokens: rate applies to the entire request; Cache read / cache write: 2.2 / 27.5 USD; Astra example prices distinguish Global and US Data Zone, and short and long context |
| GPT-6 Astra Provisioned Throughput | contact sales | usage-based | Alternative to Standard; booked capacity is charged even without calls; Hourly billing or monthly/annual reservation; Astra PTU amounts are not listed; The Astra announcement gives US Data Zone a 10% premium over Global PTU prices; confirm a specific quote |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://azure.microsoft.com/en-us/products/ai-foundry
- https://azure.microsoft.com/en-us/blog/gpt-6-astra-frontier-intelligence-for-work-now-generally-available-in-microsoft-foundry/
- https://learn.microsoft.com/en-us/azure/foundry/concepts/manage-costs
- https://learn.microsoft.com/en-us/azure/foundry/foundry-models/concepts/models-sold-directly-by-azure
- https://learn.microsoft.com/en-us/azure/foundry/responsible-ai/openai/data-privacy
- https://azure.microsoft.com/en-us/pricing/purchase-options/azure-account
- https://devblogs.microsoft.com/foundry/whats-new-in-microsoft-foundry-oct-nov-2025/
- https://azure.microsoft.com/en-us/pricing/details/azure-openai/

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/azure-ai-foundry*