Registry / Mistral via Azure AI
Mistral via Azure AI
Mistral AI models served on Microsoft Azure via Azure AI Foundry (Microsoft Foundry) Models. Some Mistral models (e.g. Mistral Large 3, Mistral Document AI) are "sold directly by Azure", hosted and operated by Azure under Azure SLAs, while others are offered as serverless Models-as-a-Service from the partners-and-community collection, where the model is a Non-Microsoft Product but Microsoft manages the hosting and acts as data processor.
Watch-outs 5
The cells where this offering is not a clean public yes. This is what to check before you sign.
- SOC 2 Type II: Partial SOC 2 report covers Azure platform services only; third‑party Mistral model is not in scope
- ISO 27001: Yes, sales-gated certification report requires request via Trust Center
- ISO 42001: Partial The certification is Microsoft's platform-level AIMS for Azure AI Foundry Models; Mistral AI itself does not publicly hold an ISO/IEC 42001 certificate, so the model is covered only via the platform's scope, not individually.
- HIPAA BAA: Partial Microsoft publicly offers a BAA by default covering in-scope Azure services, but Microsoft publishes no explicit confirmation that serverless Mistral models in Azure AI Foundry are individually in BAA scope.
- Retention / ZDR: Partial Zero‑data‑retention (ZDR) requires a gated request via Modified Abuse Monitoring.
Azure provides a publicly gated SOC 2 Type II report for its platform, but the report’s scope lists only Azure services and does not include the Mistral model, so coverage is partial.
You can access Azure SOC audit reports and bridge letters from the Service Trust Portal (STP) [SOC reports] section. Services in scope - Azure - Dynamics 365 - Microsoft 365 - Power Platform
Mistral's own ISO/IEC 27001 certification exists for the model, but the actual report is gated, so the offering is covered under yes_sales_gated.
Yes, Mistral complies with both ... ISO 27001/27701 frameworks. ... for more information, and to request a copy of our Compliance Reports, please visit our Trust Center.
Microsoft's own blog confirms the Azure AI Foundry Models ISO/IEC 42001:2023 certification, but Mistral AI's trust center (trust.mistral.ai/resources) lists only ISO 27001, ISO 27701, SOC 2 and HIPAA BAA with no ISO 42001, so per the model-developer-first rule the value stays partial.
Microsoft has achieved ISO/IEC 42001:2023 certification—a globally recognized standard for Artificial Intelligence Management Systems (AIMS) for both Azure AI Foundry Models and Microsoft Security Copilot.
The Microsoft Service Trust Portal provides a publicly accessible AI Resources page, serving as a maintained compliance portal for Azure AI offerings including Mistral via Azure AI.
Resources describing the approach to Compliance, Security and Privacy in the Artificial Intelligence (AI) solutions such as Copilot and Azure Open AI
no_verified is contradicted by Microsoft's own pages (learn.microsoft.com/en-us/compliance/regulatory/offering-hipaa-hitech: "Can my organization enter into a BAA with Microsoft? Yes...covers in-scope Microsoft services" with "Azure and Azure Government" listed, and learn.microsoft.com/en-us/azure/c
"Can my organization enter into a BAA with Microsoft? Yes. Microsoft offers its covered entity and business associate customers a Business Associate Agreement that covers in-scope Microsoft services." — in-scope list includes "Azure and Azure Government"; Azure HIPAA page adds: "Microsoft will enter into BAAs with its covered entity and business associate customers ... offers a HIPAA BAA as part of the Microsoft Product Terms ... to all customers who are covered entities or business associates under HIPAA for use of such in-scope Azure services."
Microsoft provides a publicly downloadable DPA that implements the EU Standard Contractual Clauses and a publicly accessible subprocessor list covering services under that DPA.
released an updated version of the Microsoft Products and Services Data Protection Addendum (DPA) implementing the new SCC...; Microsoft Online Services Subprocessors ... identifies subprocessors authorized to subprocess customer or personal data in Microsoft Online Services. This list is applicable for all Microsoft Online Services governed by the Microsoft Data Protection Addendum.
The Azure Foundry data‑privacy page publicly states that prompts and outputs are not used to train or improve provider models, applying to Mistral as a model sold by Azure.
Your prompts (inputs) and completions (outputs), your embeddings, and your training data: ... are NOT used by providers of Models sold by Azure to improve their models or services. ... are NOT used to train any generative AI foundation models without your permission or instruction.
Retention policy is publicly documented, but zero‑data‑retention is only available after an application, so only part of the claim is publicly accessible.
The models are stateless: no prompts or completions are stored in the model.
Microsoft's own docs (learn.microsoft.com deployment-types: "Data Zone types: The service processes data only within the Microsoft-specified data zone (US, EU, or Asia Pacific)"; EU Data Zone follows the Azure EU Data Boundary) and its Foundry blog confirm Mistral models on Azure support EU Data Zon
Mistral Large 4 will be available in Sweden Central region for Global Standard and EU DataZone deployment, providing options for teams with data residency requirements.
The EU Commission's signatory list page includes Mistral AI, confirming the developer is a signatory.
- Mistral AI
The developer (Mistral AI) publishes the Art. 53 training-content summary for Mistral Large 3 on its own legal.mistral.ai domain following the EC template, with no gate, satisfying the yes_public criterion.
Public Summary of Training Content
Spotted an error? Submit a correction with evidence, corrections with a primary source are folded in and credited in the changelog.