--- title: "Mistral Small 4 | Ethical Stack" url: https://ethicalstack.ai/models/mistral-small-4/ catalog_version: f6baff2119d0e9158520da7fbe4232011dac79f25bd08757436d7a1dbd593db2 --- # Mistral Small 4 119B mixture-of-experts model with 6.5B active parameters, text-and-image input, reasoning modes, function calling and a 256k context. Canonical profile: https://ethicalstack.ai/models/mistral-small-4/ Scope: Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate. Published overall: 3.2; 5/6 dimensions rated. Round the equal-weight mean of available numerical ratings to one decimal; require at least four of six for an overall score. BDS status stays outside the mean; for eligible profiles, a boycott target caps the final score at 2/10 and a pressure target subtracts one point (minimum 1). Not listed and not checked do not adjust the score. ## Energy practices Finding: 3/10 Assessment: https://ethicalstack.ai/models/mistral-small-4/#energy This grades disclosed operating practices, not the electricity or emissions of a particular generation. Reason: 1 + 1 reporting + 0 electricity + 1 controls = 3/10. This grades documented practices and transparency. Argument: The release gives limited runtime or control details but no comparable per-task footprint or electricity-source evidence. Mistral documents a local inference boundary and recommends hardware for a 119B checkpoint. Mistral’s offline guide says to reduce context length or use 4-bit precision for lower-memory configurations; these are documented runtime controls, not measured energy savings. Type: argument; confidence: medium Scope (dimension): Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate. Research reviewed: not recorded; evidence date: not recorded Unknowns: Per-task energy, serving region, facility electricity mix, water and independent footprint review are not established in the cited release sources. ```json { "method": "practices-transparency-v1", "scope": "Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate.", "components": { "disclosure": { "score": 1, "reason": "Mistral documents a local inference boundary and recommends hardware for a 119B checkpoint.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] }, "electricity": { "score": 0, "reason": "The release materials do not report matched regional electricity sourcing or carbon-free supply for training or inference.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] }, "demand": { "score": 1, "reason": "Mistral’s offline guide says to reduce context length or use 4-bit precision for lower-memory configurations; these are documented runtime controls, not measured energy savings.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] } }, "actualImpactMeasured": false } ``` - [Mistral Small 4 119B A6B official model card](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) (primary; checked 2026-10-24) - [Mistral Small 4 official API model page](https://docs.mistral.ai/models/mistral-small-4-0-26-03) (primary; checked 2026-10-24) - [Mistral offline-model deployment documentation](https://docs.mistral.ai/vibe/code/cli/offline-models) (primary; checked 2026-10-24) - [Mistral privacy policy](https://legal.mistral.ai/terms/privacy-policy/) (primary; checked 2026-10-24) ## Training data Finding: 1/10 Assessment: https://ethicalstack.ai/models/mistral-small-4/#training-rights The model release does not provide a complete inspectable training-source permission record. Reason: 1 base + 0 source transparency + 0 documented permission + 0 creator controls = 1/10. Argument: The model card describes architecture, modalities and license but does not identify training data sources or permissions. It provides no source-level permission ledger or creator remedy; a model-weight license governs downstream use, not the material used to train the model. Type: argument; confidence: medium Scope (profile): Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate. Research reviewed: not recorded; evidence date: not recorded Unknowns: Named source inventory, source-level licenses, creator remedies, and independent verification remain unestablished for this release. ```json { "method": "training-rights-evidence-v1", "scope": "The stated training release or selected model", "components": { "provenance": { "score": 0, "reason": "The model card describes architecture, modalities and license but does not identify training data sources or permissions.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] }, "permissions": { "score": 0, "reason": "The release sources do not publish a training-source permission basis or rights ledger; the weight license does not establish training-data permission.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] }, "creatorControls": { "score": 0, "reason": "No training-source creator permission-before-use, opt-out, removal or retraining route is documented for this release.", "sourceUrls": [ "https://huggingface.co/mistralai/Mistral-Small-4-119B-2603", "https://docs.mistral.ai/models/mistral-small-4-0-26-03", "https://docs.mistral.ai/vibe/code/cli/offline-models", "https://legal.mistral.ai/terms/privacy-policy/" ] } }, "routeDependency": null, "modelReference": null } ``` - [Mistral Small 4 119B A6B official model card](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) (primary; checked 2026-10-24) - [Mistral Small 4 official API model page](https://docs.mistral.ai/models/mistral-small-4-0-26-03) (primary; checked 2026-10-24) - [Mistral offline-model deployment documentation](https://docs.mistral.ai/vibe/code/cli/offline-models) (primary; checked 2026-10-24) - [Mistral privacy policy](https://legal.mistral.ai/terms/privacy-policy/) (primary; checked 2026-10-24) ## Your data & privacy Finding: Not enough evidence Assessment: https://ethicalstack.ai/models/mistral-small-4/#privacy Privacy has no single score for these downloadable weights because handling depends on the inference host and integrations. Reason: Depends on selected inference host; no single prompt-handling route applies to the weights. Argument: The release can be run locally, but it also appears in hosted provider routes; those operators set their own request retention, training and region terms. A cloud API, where offered, is a separate policy route and is not inherited from the local weights. Type: argument; confidence: high Scope (dimension): Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate. Research reviewed: not recorded; evidence date: not recorded Unknowns: A selected host, network path, telemetry, logging, external prompt enhancement, and data retention need route-specific review. - [Mistral Small 4 119B A6B official model card](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) (primary; checked 2026-10-24) - [Mistral Small 4 official API model page](https://docs.mistral.ai/models/mistral-small-4-0-26-03) (primary; checked 2026-10-24) - [Mistral offline-model deployment documentation](https://docs.mistral.ai/vibe/code/cli/offline-models) (primary; checked 2026-10-24) - [Mistral privacy policy](https://legal.mistral.ai/terms/privacy-policy/) (primary; checked 2026-10-24) ## Openness Finding: 7/10 Assessment: https://ethicalstack.ai/models/mistral-small-4/#openness Apache 2.0 weights and inference instructions are public, but no release-specific full training recipe or corpus is published. Reason: 7/10 reflects permissive weights and reproducible inference, with training data and full training recipe unavailable. Argument: The model card provides downloadable weights, an Apache 2.0 license and executable Transformers/vLLM instructions. It describes the architecture, but not a complete training recipe or dataset inventory. This rating concerns release artifacts and reproducibility, not training-data permission. Type: argument; confidence: medium Scope (profile): Mistral Small 4 release v26.03 weights and Mistral-hosted API model mistral-small-2603; local and hosted handling are separate. Research reviewed: not recorded; evidence date: not recorded Unknowns: Training data provenance and the full pretraining/post-training recipe are not disclosed for this release. - [Mistral Small 4 119B A6B official model card](https://huggingface.co/mistralai/Mistral-Small-4-119B-2603) (primary; checked 2026-10-24) - [Mistral Small 4 official API model page](https://docs.mistral.ai/models/mistral-small-4-0-26-03) (primary; checked 2026-10-24) - [Mistral offline-model deployment documentation](https://docs.mistral.ai/vibe/code/cli/offline-models) (primary; checked 2026-10-24) - [Mistral privacy policy](https://legal.mistral.ai/terms/privacy-policy/) (primary; checked 2026-10-24) ## Worker protections Finding: 2/10 Assessment: https://ethicalstack.ai/models/mistral-small-4/#labour Mistral publicly says its supplier code applies to goods and services and its Trust Center shows contractors acknowledge a code, but labor-specific provisions and outcomes were not publicly inspectable. Reason: 1 + 1 pay & conditions + 0 worker voice & remedy + 0 evidence & follow-through = 2/10. This grades documented worker safeguards, not a measured rate of mistreatment. Argument: Mistral’s August 2026 help page states that its Code of Conduct sets standards suppliers are expected to follow for Mistral business and all supplier goods/services; the Trust Center lists contractor acknowledgment of the Code of Conduct. The public interface did not expose the code’s labor clauses or worker procedures, and no worker survey, audit, pay/conditions disclosure or remediation results were located. This supports limited credit for a supplier-code commitment only. This entry carries the same developer/operator finding to Mistral Small 4; it does not establish that this product was deployed under every cited relationship. Type: argument; confidence: low Scope (dimension): Mistral AI: the documented developer/operator record, not proof of deployment of Mistral Small 4. Research reviewed: 2026-10-24; evidence date: not recorded Unknowns: The supplier code’s detailed labor language and coverage of AI data workers, grievance route, monitoring and outcomes were not accessible from the public Trust Center. ```json { "method": "worker-safeguards-evidence-v1", "entity": "mistral", "company": "Mistral AI", "scope": "Mistral AI's workforce and disclosed contractors. Coverage of individual product teams may be incomplete.", "components": { "conditions": { "score": 1, "reason": "A supplier code is a stated business requirement, but its labor content and implementation were not publicly inspectable in this review.", "sourceUrls": [ "https://help.mistral.ai/en/articles/347637-do-you-have-a-code-of-conduct", "https://trust.mistral.ai/" ] }, "voice": { "score": 0, "reason": "No public worker-specific voice or grievance mechanism was established in the accessible sources.", "sourceUrls": [ "https://help.mistral.ai/en/articles/347637-do-you-have-a-code-of-conduct", "https://trust.mistral.ai/" ] }, "assurance": { "score": 0, "reason": "The Trust Center shows contractor code acknowledgment but no labor-specific audits, remediation or outcomes.", "sourceUrls": [ "https://help.mistral.ai/en/articles/347637-do-you-have-a-code-of-conduct", "https://trust.mistral.ai/" ] } }, "disclosureLimited": true, "meaning": "This grades documented worker safeguards, not a measured rate of mistreatment." } ``` - [Do you have a Code of Conduct?](https://help.mistral.ai/en/articles/347637-do-you-have-a-code-of-conduct) (primary; checked 2026-10-24) - [Mistral AI Trust Center](https://trust.mistral.ai/) (primary; checked 2026-10-24) ## Military ties Finding: 3/10 Assessment: https://ethicalstack.ai/models/mistral-small-4/#military-contracts Mistral AI publicly partners with defense AI company Helsing on joint development of vision-language-action models, and its current company page says it serves defense customers. A separate Airbus agreement includes military aerospace applications; specific Mistral model deployments in weapons systems are not established. Reason: 1 + 0 use restrictions + 0 enforcement & oversight + 2 relationship disclosure = 3/10. Known defense ties lower this rating; safeguards can improve it within the relationship’s cap. Argument: Mistral's announcement with Helsing describes joint development of next-generation AI systems for European defense, while Helsing's announcement links its military AI portfolio to Ukraine-deployed strike drones and fighter electronic-warfare capabilities. Mistral also identifies defense as a customer sector and Airbus says the partnership includes military aerospace applications. This is broad defense enablement, capped at 3/10; no binding weapons/surveillance prohibitions or military-use oversight outcomes are public in these sources. This entry carries the same developer/operator finding to Mistral Small 4; it does not establish that this product was deployed under every cited relationship. Type: argument; confidence: medium Scope (dimension): Mistral AI: the documented developer/operator record, not proof of deployment of Mistral Small 4. Research reviewed: 2026-10-24; evidence date: not recorded Unknowns: The sources do not establish which Mistral models, if any, are deployed in specific military systems, the operational purpose or safeguards, contract restrictions, or independent oversight. Helsing's existing strike systems are not evidence that Mistral models themselves control or target weapons. ```json { "method": "military-ties-priority-v2", "entity": "mistral", "company": "Mistral AI", "scope": "Mistral AI's relationships and safeguards. A company relationship does not identify every product or release used.", "components": { "restrictions": { "score": 0, "reason": "The reviewed public partnership and company materials do not disclose binding restrictions on weapons, harmful surveillance, or military-intelligence use.", "sourceUrls": [ "https://helsing.ai/newsroom/helsing-and-mistral-announce-strategic-partnership-in-defence-ai", "https://mistral.ai/about/" ] }, "oversight": { "score": 0, "reason": "No public military-specific review, enforcement mechanism, or independent outcome assessment was identified in the reviewed partnership materials.", "sourceUrls": [ "https://helsing.ai/newsroom/helsing-and-mistral-announce-strategic-partnership-in-defence-ai", "https://mistral.ai/about/" ] }, "transparency": { "score": 2, "reason": "Mistral and its partners identify the defense relationships and intended applications, including VLA model development and military aerospace, though no model-level deployment details or outcomes are disclosed.", "sourceUrls": [ "https://helsing.ai/newsroom/helsing-and-mistral-announce-strategic-partnership-in-defence-ai", "https://www.airbus.com/en/newsroom/press-releases/2026-05-airbus-partners-with-mistral-ai-to-strengthen-the-use-of-artificial-intelligence-in-sovereign" ] } }, "relationship": "broad-defense", "relationshipReason": "Mistral publicly confirms joint defense-AI development and military-aerospace partnerships, but no source establishes its model in actual targeting or combat intelligence use.", "score": 3, "baseScore": 1, "safeguardPoints": 2, "safeguardAdjustment": 2, "rawScore": 3, "cap": 3, "disclosureLimited": true, "meaning": "Known military relationships take priority over missing safeguards. The rating does not establish the probability of an undisclosed contract." } ``` - [Helsing and Mistral announce strategic partnership in defence AI](https://helsing.ai/newsroom/helsing-and-mistral-announce-strategic-partnership-in-defence-ai) (primary; checked 2026-10-24) - [About Mistral](https://mistral.ai/about/) (primary; checked 2026-10-24) - [Airbus and Mistral AI partner for AI in aerospace](https://www.airbus.com/en/newsroom/press-releases/2026-05-airbus-partners-with-mistral-ai-to-strengthen-the-use-of-artificial-intelligence-in-sovereign) (primary; checked 2026-10-24) ## BDS targets Finding: Not listed Assessment: https://ethicalstack.ai/models/mistral-small-4/#israel-palestine We did not find Mistral AI among the named targets in the two official BNC campaign lists checked on 2026-10-23. This is a limited list check, not BDS approval. Reason: BDS campaign status is a factual category, not a 1–10 ethical rating. Argument: Mistral AI is not a named target in the checked lists, but BDS uses selective campaigns, so absence cannot establish a clean record. Running downloaded weights locally is a different decision from paying the developer or a cloud host; check the host as well. This entry carries the same developer/operator finding to Mistral Small 4; it does not establish that this product was deployed under every cited relationship. Type: evidence; confidence: medium Scope (dimension): Mistral AI: the documented developer/operator record, not proof of deployment of Mistral Small 4. Research reviewed: 2026-10-23; evidence date: not recorded Unknowns: This review covers the named targets in the BNC consumer/pressure guide and No Tech campaign, not every BDS campaign, local group, investment or supplier. Lists and ownership can change. A company’s absence from these lists is not an endorsement or proof that it has no relevant relationships. - [BNC: live guide to consumer boycott and pressure targets](https://bdsmovement.net/Guide-to-BDS-Boycott) (campaign; checked 2026-10-23) - [BNC: No Tech for Oppression, Apartheid or Genocide campaign](https://bdsmovement.net/no-tech-oppression-apartheid-or-genocide) (campaign; checked 2026-10-23) ## Reading the evidence - Assessments are provisional editorial judgments. Cite the profile and the underlying sources, and retain confidence, scope and unknowns. - Unknown, not-applicable and model-dependent are different states. Missing evidence does not establish good or bad conduct. - Company-level evidence does not establish use of a particular product or model. Account plan, model, host and route can change an assessment. - Energy scores describe documented practices and transparency, not measured request emissions. Zero component points can mean an evidence gap. - Published overall scores require at least four of six numerical dimensions. Preserve coverage and BDS adjustments; BDS itself is categorical. - No named BDS target found is limited to the reviewed campaign sources and is not approval. Personalized priorities do not change the published overall score. - Review and source-check dates describe the research, not this deployment. Fetch current primary sources when the decision requires current verification. Methodology: https://ethicalstack.ai/methodology/