The model card discloses that Mistral Medium 3.5 accepts text and image inputs, produces text outputs, supports a 128,000-token context window, and is optimized for agentic workflows and coding tasks.
This analysis describes what Mistral AI's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology
These technical specifications are operationally significant for deployers evaluating API pricing, infrastructure planning, and use-case suitability; the 128k context window and multimodal (vision) input capability determine which application architectures are feasible and what data categories may be processed by the model.
The disclosed specifications establish that the model can process image inputs in addition to text, and that users or deployers may submit up to 128,000 tokens per context, which is relevant to understanding what types of content and how much data may be submitted to the model in a single interaction.
Cross-platform context
See how other platforms handle Model Technical Specifications and Context Window and similar clauses.
Compare across platforms →Monitoring
Mistral AI has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"Our frontier-class multimodal model optimized for agentic and coding use cases. Released as open weights under a Modified MIT license.— Excerpt from Mistral AI's Mistral Medium 3.5 Model Card
1. REGULATORY LANDSCAPE: The acceptance of image inputs in addition to text means that data processing under this model may involve biometric or sensitive visual information depending on the images submitted, engaging GDPR Article 9 (special categories of personal data) for EU deployers and potentially CCPA for California-based users if images contain personal information. The FTC's guidance on AI and consumer data practices is also relevant for US-based deployers processing consumer-supplied images. 2. GOVERNANCE EXPOSURE: Medium. Deployers who submit user-uploaded images to the model API must ensure their data processing agreements and privacy notices account for the multimodal input capability. Organizations that assumed only text data was processed may need to update data mapping and DPIA documentation. 3. JURISDICTION FLAGS: EU and EEA deployers processing images that may contain personal data (including faces or identifiable information) face GDPR obligations including lawful basis documentation and potential DPIA requirements. Illinois BIPA may apply if images containing biometric identifiers are processed. California CCPA/CPRA applies to personal information in images submitted by California consumers. 4. CONTRACT AND VENDOR IMPLICATIONS: Data processing agreements with Mistral AI should explicitly address the multimodal input capability and confirm how image data submitted via the API is handled, retained, and whether it is used for model training. Deployers should confirm these terms before enabling image input in user-facing products. 5. COMPLIANCE CONSIDERATIONS: Organizations should: (a) update data flow maps to include image data submitted to the Mistral API; (b) review privacy notices to ensure multimodal data processing is disclosed to end users; (c) assess whether DPIAs are required for image processing use cases; (d) confirm with Mistral AI the data retention and training use terms for image inputs specifically.
Full compliance analysis
Regulatory citations, enforcement risk, and due diligence action items.
Free: track 3 platforms + weekly digest. Monitor: 25 platforms + same-day alerts. No credit card required.
Compliance Governance Intelligence
Need to monitor specific governance provisions?
Compliance includes provision-level monitoring, governance timelines, regulatory mapping, and audit-ready analysis.
Built from archived source documents, structured governance mappings, and historical version tracking.
These technical specifications are operationally significant for deployers evaluating API pricing, infrastructure planning, and use-case suitability; the 128k context window and multimodal (vision) input capability determine which application architectures are feasible and what data categories may be processed by the model.
The disclosed specifications establish that the model can process image inputs in addition to text, and that users or deployers may submit up to 128,000 tokens per context, which is relevant to understanding what types of content and how much data may be submitted to the model in a single interaction.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Mistral AI.