Additional Product Terms for Use of Mistral AI Products on Mistral AI Infrastructure

These Additional Terms for Use of Mistral AI Products on Mistral AI Infrastructure supplements the Terms where Customer is authorized to access and use Mistral AI Products on Mistral AI Infrastructure.

1. Usage Policy. When Customer uses the Mistral AI Products on the Mistral AI Infrastructure, Customer must comply with Mistral AI's Usage Policy.

2. Customer and End User Accounts. Use of Mistral AI Products on Mistral AI Infrastructure requires a Customer Account. Customer must provide complete and accurate account information when setting up its Customer Account, including accurate billing information and a valid method of payment. Subject to applicable limitations (e.g., a maximum number of seat licenses), Customer may create End User Accounts under its Customer Account to enable End Users to receive direct access to the Mistral AI Products. Customer must ensure that all account details for its Customer Account and all End User Accounts remain accurate and all account credentials remain confidential at all times. End User Accounts may not be shared, sold, licensed, or otherwise made available outside of Customer’s entity or organization. For the avoidance of doubt, but subject to Section 1.5 (Customer Affiliates) of the Terms, Customer may create End User Accounts for End Users of its Affiliates. Customer must promptly notify Mistral AI at support@mistral.ai or through the Help Center in the event Customer detects any unauthorized or fraudulent access or use of its Customer Account, any End User Account, or any Customer Offering.

3. Media Customers. “Media Customer” means any entity that disseminates information for editorial purposes on a regular basis and in a professional capacity, whether to the general public or to professionals, even if such activity constitutes only a part of its overall business activities. If Customer is a Media Customer, Customer is not authorized to use the verified news feature, such as the AFP news feature, in the Mistral AI Products for editorial use. Any use by Customer of such news feature may be automatically blocked in Mistral AI’s sole discretion and without refund. If Customer is a Media Customer and wishes to use the news feature for editorial purposes, Customer must contact Mistral AI at legal@mistral.ai. Any unauthorized use of the news feature by a Media Customer will result in the immediate suspension of such Media Customer’s account.

4. Sharing Conversation Links. Certain Mistral AI Products provide users with the ability to share unique links to conversations with third parties (each, a “Conversation Link”). Anyone with the Conversation Link will be able to view Customer’s conversation, including any person with whom Customer’s intended recipient shares the Conversation Link. Customer acknowledges that Mistral AI does not monitor or control who accesses conversations via Conversation Links that Customer or an End User creates.

5. Third Party Services. Customer may connect services provided by third parties (each, a “Third-Party Service”), including Model Context Protocol servers (“MCP Servers”), to Mistral AI Products. By enabling these connections, Customer grants Mistral AI permission to (a) access the Third-Party Services, (b) send Customer Inputs to the Third-Party Service, and (c) where applicable, retrieve, process, and store the necessary data to use as Inputs to generate Outputs. Customer may only connect a Third-Party Service to a Mistral AI Product if Customer has all necessary rights, licenses, and permissions to connect such Third-Party Service to Mistral AI Products in accordance with these Terms. Customer is responsible for maintaining access to the Third-Party Service for the duration of Customer’s subscription or Customer’s use of such features. Any use of a Third-Party Service is subject to the respective terms and conditions for the Third-Party Service, and Customer is solely responsible for complying with such terms and conditions. Customer acknowledges that Third-Party Services are not Mistral AI Products and Mistral AI is not responsible for any Third-Party Service or Customer use of any Third-Party Service, including any disruptions, data loss, or issues with Mistral AI Products arising from Customer’s use of the Third-Party Service.

6. Moderation. Mistral AI may monitor use of the Mistral AI Products through automated means in accordance with the Usage Policy. This monitoring is conducted to ensure compliance with Mistral AI’s terms and policies, and to maintain the security and integrity of Mistral AI Products.

7. Third-Party Content. When Customer uses certain features on a Mistral AI Product, such as web-search, Customer may receive (a) Output, and (b) content provided by third-parties in the form of additional resources such as hyperlinks, images, thumbnails, or snippets from third-parties to help you further explore or understand the Output (the “Third-Party Content”). Such Third-Party Content does not constitute Customer Content and Mistral AI does not grant you any rights to this Third-Party Content, other than viewing it. For clarity, Mistral AI does not use the Third-Party Content provided in Customer's Input or that is displayed in Outputs to train our artificial intelligence models. Customer is not allowed to use Third-Party Content displayed using Mistral AI's web-search feature to (i) copy, store, archive, cache or create a database of the Third-Party Content, (ii) redistribute, resell, or sublicense the Third-Party Content, (iii) as part of any machine learning or similar algorithmic activity, or (iv) to create, train, evaluate or improve commercial products or services that you make available to third-parties.

8. Prohibited Content. Customer must promptly report to Mistral AI any Customer Content or Third-Party Content that Customer identifies at any time as violating Mistral AI’s Usage Policy or which otherwise (a) incites hate, violence, or discrimination against individuals based on their origin, ethnicity, religion, gender, sexual orientation, or similar characteristics, (b) glorifies sexual harassment, (c) violates human dignity (e.g., human trafficking), (d) glorifies crimes against humanity or denying their existence, (e) incites terrorism, (f) glorifies severe crimes against individuals (e.g., murder or sexual assault), (g) glorifies crimes involving theft, extortion, or material damage posing a danger to individuals, (h) constitutes child sexual abuse material, or (i) is dangerous for minors. Customer can report such content by using the report feature in the Mistral AI Products or by sending an email to support@mistral.ai. Customer grants Mistral AI the worldwide, perpetual, irrevocable right to access and use Customer Content associated with the reported content to improve the Mistral AI Products (e.g., to help us prevent the Mistral AI Products from generating such content again).

Additional Terms for Use of Mistral AI Products on Customer Infrastructure

These Additional Terms for Use of Mistral AI Products on Customer Infrastructure supplements the Terms where Customer is authorized to access and use Mistral AI Products on Customer Infrastructure, as provided in an Order Form between Customer and Mistral AI.

1. Technical Requirements. The Documentation provides Mistral AI’s general recommendations regarding the configuration of Customer Infrastructure for use of the Mistral AI Products (“Configuration Recommendations”). Customer acknowledges that the Configuration Recommendations are recommendations only and Customer is solely responsible for configuring Customer Infrastructure for the deployment of Mistral AI Products and fixing any issues associated with Customer Infrastructure. For the avoidance of doubt, Mistral AI is not responsible for any interruptions, delays, data loss, service failures, or other problems that may occur when using the Mistral AI Products on Customer Infrastructure.

2. Reporting. Where Customer subscribes to Mistral AI Products with a usage limitation (e.g., seats, pages, nodes) (each, a “Usage Metric”), Customer will provide reports to Mistral AI on a quarterly basis (on January 31, April 30, July 31, October 31) no later than the fifth (5th) day after the end of each quarter]. Such reports will be sent to ar@mistral.ai and include a detailed count of the Usage Metric during the preceding quarter.

3. Audit. While Customer’s Terms remain in effect and for one (1) year thereafter, Mistral AI may, at its own expense and with ten (10) business days’ notice to Customer, conduct audits (up to one (1) audit per year) to ensure Customer’s compliance with the Terms (including any Additional Terms). Each audit may be conducted by Mistral AI or a third-party auditor of Mistral AI’s choosing. Customer agrees to cooperate in good faith and provide all relevant documentation and grant access to its premises and computer systems to enable Mistral AI to assess Customer’s compliance. If the audit reveals unauthorized use of the Mistral AI Products or Confidential Information of Mistral AI, Mistral AI may: (a) charge Customer additional fees that Customer should have paid due to such unauthorized use; or (b) mandate the destruction of unauthorized versions or copies of the Mistral AI Products or Confidential Information of Mistral AI. In the event of audit findings indicating Customer’s breach of the Agreement, the costs of the audit will be borne by Customer only. Mistral AI’s exercise of its rights under this Section 3 (Audit) will be without prejudice to any other remedies Mistral AI may have under the Terms or under applicable law.

4. Security Incidents. In the event of any suspected or actual unauthorized disclosure or access to the Mistral AI Products or their weights (each, a “Security Incident”), Customer will immediately notify Mistral AI at legal@mistral.ai. Such notification must include, at minimum, the following details: (a) the time the Security Incident occurred, (b) the Mistral AI Product affected by the Security Incident, (c) the nature and impact of the Security Incident, and (d) the measures which have been taken or which are proposed to be taken to address the Security Incident and to mitigate its possible adverse effects. Customer agrees to take those steps reasonably requested by Mistral AI (at Customer’s sole expense) to minimize damage to Mistral AI in connection with such a Security Incident.

5. Media Customers. “Media Customer” means any entity that disseminates information for editorial purposes on a regular basis and in a professional capacity, whether to the general public or to professionals, even if such activity constitutes only a part of its overall business activities. If Customer is a Media Customer, Customer is not authorized to use the verified news feature, such as the AFP news feature, in the Mistral AI Products for editorial use. Any use by Customer of such news feature may be automatically blocked in Mistral AI’s sole discretion and without refund. If Customer is a Media Customer and wishes to use the news feature for editorial purposes, Customer must contact Mistral AI at legal@mistral.ai. Any unauthorized use of the news feature by a Media Customer will result in the immediate suspension of such Media Customer’s account.

6. Sharing Conversation Links. Certain Mistral AI Products provide users with the ability to share unique links to conversations with third parties (each, a “Conversation Link”). Anyone with the Conversation Link will be able to view Customer’s conversation, including any person with whom Customer’s intended recipient shares the Conversation Link. Customer acknowledges that Mistral AI does not monitor or control who accesses conversations via Conversation Links that Customer or an End User creates.

7. Third Party Services. Customer may connect services provided by third parties (each, a “Third-Party Service”), including Model Context Protocol servers (“MCP Servers”), to Mistral AI Products. By enabling these connections, Customer grants Mistral AI permission to (a) access the Third-Party Services, (b) send Customer Inputs to the Third-Party Service, and (c) where applicable, retrieve, process, and store the necessary data to use as Inputs to generate Outputs. Customer may only connect a Third-Party Service to a Mistral AI Product if Customer has all necessary rights, licenses, and permissions to connect such Third-Party Service to Mistral AI Products in accordance with these Terms. Customer is responsible for maintaining access to the Third-Party Service for the duration of Customer’s subscription or Customer’s use of such features. Any use of a Third-Party Service is subject to the respective terms and conditions for the Third-Party Service, and Customer is solely responsible for complying with such terms and conditions. Customer acknowledges that Third-Party Services are not Mistral AI Products and Mistral AI is not responsible for any Third-Party Service or Customer use of any Third-Party Service, including any disruptions, data loss, or issues with Mistral AI Products arising from Customer’s use of the Third-Party Service.

8. Survival. Section 3 (Audit) and Section 4 (Security Incidents) will survive any termination of the Terms.

AI Laws Responsibilities

These Additional Terms for AI Laws Responsibilities form an integral part of the agreement between you ("Customer") and Mistral AI (the “Agreement”). They outlines each parties’ obligations under the Applicable AI Laws.

1. Definitions

“Applicable AI Laws” means any applicable artificial intelligence law or regulation, including, to the extent applicable (a) Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonized rules on artificial intelligence including any associated implementation or delegated acts (the “AI Act”), and (b) mandatory guidelines published by a Competent Authority.
“AI Act Website” means the interface provided by Mistral AI on its website where Mistral AI makes available to Customer (a) the list of Covered Mistral AI Products and (b) the Downstream Provider Documentation.
“Competent Authority” means a governmental authority with competent jurisdiction over the enforcement of an Applicable AI Law, including with respect to the AI Act (a) the AI Office and (b) the European Artificial Intelligence Board.
“Covered Customer Product” means any product provided by Customer to any End-User, including any third-party, as authorized under this Agreement that (a) integrates a Mistral AI Product, and/or results from a modification of a Mistral AI Product by or on behalf of Customer, and (b) is subject to the Applicable AI Laws. Covered Customer Products may include any (i) General-Purpose AI Model, (ii) Open-Source General-Purpose AI Model, (iii) General-Purpose AI Model with systemic risks, (iv) AI System or (v) High-Risk AI System.
“Covered Mistral AI Product” means any Mistral AI Product provided by Mistral AI to Customer under this Agreement and that is subject to the Applicable AI Laws, including any (a) General-Purpose AI Model, (b) Open-Source General-Purpose AI Model, (c) General-Purpose AI Model with systemic risks or (d) AI System.
“Covered Third-Party Products” means any product provided by a third-party under this Agreement and that is subject to the Applicable AI Laws including any (a) General-Purpose AI Model, (b) Open-Source General-Purpose AI Model, (c) General-Purpose AI Model with systemic risks, (c) AI System or (d) High-Risk AI System.
“Downstream Provider Documentation” means the documentation that Mistral AI is required to make available to Customer as Downstream Provider under the Applicable AI Laws, including for the avoidance of doubt the information and documentation regarding the capabilities and limitations of a General-Purpose AI Model or a General-Purpose AI Model with systemic risks provided by Mistral AI to Customer, as required by Article 53.1 (b) of the AI Act or the analogous provision in the Applicable AI Law (if any).
“Open-Source General-Purpose AI Model” means a General Purpose AI Model released under a free and open-source licence that allows for the access, usage, modification, and distribution of the Model, and whose parameters, including the weights, the information on the model architecture, and the information on Model usage, are made publicly available, in accordance with Article 53.2 of the AI Act or the analogous provision in the Applicable AI Law (if any).

All capitalized terms under these Additional Terms, unless otherwise defined under this Agreement, shall have the meaning given to them under the Applicable AI Laws.

2. Classification

2.1. Mistral AI Classification. Mistral AI is the Provider of Covered Mistral AI Products under the Applicable AI Laws. For the avoidance of doubt, Mistral AI shall not be considered as the Provider of any Covered Customer Product or Covered Third-Party Product.

2.2. Classification of the Mistral AI Products. Customer acknowledges that not all Mistral AI Products are considered as Covered Mistral AI Products under the Applicable AI Laws. Mistral AI is responsible for the classification of the Mistral AI Products in accordance with the Applicable AI Laws. For the avoidance of doubt, Mistral AI does not provide any High-Risk AI Systems under the Agreement.

3. Responsibilities

3.1. Mutual responsibilities. Each Party shall comply with their respective obligations under the Applicable AI Laws. For the avoidance of doubt, (a) Mistral AI shall comply with its obligations as Provider of the Covered Mistral AI Products, (b) Customer shall comply with its own obligations relating to the provision, deployment and/or distribution of any Covered Customer Product, as applicable.

3.2. Mistral AI responsibilities. In addition to its obligations under Section 3.1 (Mutual Responsibilities) of these Additional Terms, to the extent required under Applicable AI Laws, Mistral AI will:

Maintain and update a list of the Covered Mistral AI Products on the AI Act Website;
Indicate on the AI Act Website the classification of each Covered Mistral AI Product as either (a) a General-Purpose AI Model, (b) an Open-Source General-Purpose AI Model, (c) a General-Purpose AI Model with systemic risks, or (d) an AI System.
Make the applicable Downstream Provider Documentation available to Customers on the AI Act Website for each Covered Mistral AI Product.
Promptly inform Customer of any substantial change to the Downstream Provider Documentation.
Upon Customer’s request, Mistral AI will provide Customer with:

Additional information insofar as such information is necessary to (i) enable Customer to have a good understanding of the capabilities and limitations of the Mistral AI General-Purpose AI Model relevant for its integration into the Customer’s AI system, and (ii) enable Customer to comply with its obligations pursuant to the Applicable AI Laws. Mistral AI will provide such information within a reasonable timeframe, and no later than fourteen (14) days of receiving the request, save for exceptional circumstances, and provided that Customer’s request is sufficiently precise. Such provision is subject to Mistral AI's rights to protect its intellectual property rights, trade secrets, and confidential business information.
Reasonable information that is relevant to (i) Customer’ systemic risk assessment and mitigation where a Covered Customer Product is classified as a General-Purpose AI Model with systemic risks or (ii) Customer’s conformity assessment where Customer has repurposed an AI System provided by Mistral AI into a High-Risk AI System, in each case, in so far as necessary to enable Customer to comply with its own obligations under the Applicable AI Laws. Such provision is subject to Mistral AI's rights to protect its intellectual property rights, trade secrets, and confidential business information.

Customer acknowledges that Mistral AI may withdraw any Covered Mistral AI Product classified as a General-Purpose AI Model with systemic risks if Mistral AI reasonably determines that a systemic risk has become unacceptable. Mistral AI will give Customer reasonable prior notice before discontinuing the Covered Mistral AI Product, unless a Competent Authority requires expedited action. Mistral AI will use its commercially reasonable efforts to promptly provide Customer with an alternative Mistral AI Product.

3.3. Customer responsibilities.

If Customer uses a Covered Mistral AI Product classified by Mistral AI as General-Purpose AI Model with systemic risks, Customer acknowledges that Mistral AI is required by the Applicable AI Laws to continuously monitor and assess the systemic risks that may stem from the deployment and use of such Covered Mistral AI Product by Customer. Accordingly, upon Mistral AI’s request, Customer shall (a) reasonably cooperate with Mistral AI to implement any security or safety measure to mitigate or remediate any systemic risks in accordance with the Applicable AI Laws and (b) provide any reasonable information that is relevant to systemic risks assessment and mitigation to Mistral AI to allow Mistral AI to comply with its own obligations under the Applicable AI Laws.
Customer acknowledges that Mistral AI is responsible for reporting to the Competent Authority any Serious Incident relating to any Covered Mistral AI Product classified as a General-Purpose AI Model with systemic risks. Accordingly, Customer shall notify Mistral AI of any Serious Incident within forty-eight (48) hours of becoming aware of such Serious Incident, to allow Mistral AI to report such Serious Incident to the Competent Authority. Customer shall report the Serious Incident to Mistral AI using the form available on the AI Act Website. Customer shall not report any Serious Incident regarding a Mistral AI Product to a Competent Authority unless required to do so by the Applicable AI Laws. If required by Applicable AI Laws to report a Serious Incident to the Competent Authority, the Customer will use best efforts to notify Mistral AI prior to such reporting if the Serious Incident concerns a Covered Mistral AI Product.

4. Evolution of these Additional Terms

4.1. Evolution of these Additional Terms. The Parties acknowledge and agree that the Applicable AI Laws and especially the AI Act are new regulations that are still evolving and may impact the scope and extent of the Parties’ legal obligations. Accordingly, Mistral AI reserves the right to update these Additional Terms at any time.

4.2. Customer information. If a change to these Additional Terms has a material adverse impact on Customer, then Customer may object to the change by notifying Mistral AI within thirty (30) days after Mistral AI provides notice.

4.3. Negotiation process. If Customer notifies Mistral AI as described in Section 4.2. above, then the Parties will consult and negotiate in good faith to find a solution satisfactory for both Parties within 30 days of Customer’s notification. Failing that, Mistral AI will have the right to terminate the Agreement in accordance with the relevant provisions of the Agreement.