Get the weekly research letter
Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean. No account.
The document discloses that Command R models are optimized for 10 languages and support 13 additional languages at lower performance levels, and states that outputs in languages outside these 23 may be unreliable.
This analysis describes what Cohere'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
This disclosure establishes performance boundaries that are relevant for enterprise deployments serving multilingual populations, and creates due diligence obligations for customers deploying the models in contexts where output reliability in non-supported languages could affect user safety or regulatory compliance.
The document discloses that Command R model outputs in languages outside the 23 supported languages are described as unlikely to be reliable. For the 13 lower-resource languages within the 23, the document notes they are less rigorously evaluated than the 10 primary languages.
Cross-platform context
See how other platforms handle Language Performance Limitations Disclosure and similar clauses.
Compare across platforms →Monitoring
Cohere has changed this document before.
Receive same-day alerts, structured change summaries, and monitoring for up to 25 platforms.
"This model is designed to excel at English, French, Spanish, Italian, German, Portuguese, Japanese, Korean, Chinese, and Arabic, and to generate in 13 other languages well. It will sometimes respond in other languages, but the generations are unlikely to be reliable.Excerpt from Cohere's Responsible Use Policy
1. REGULATORY LANDSCAPE: Deployments of Command R in jurisdictions where consumer-facing communications must be accurate and in the local language, such as consumer protection requirements in the EU and various national language laws, may face compliance risk if the model's output reliability is insufficient for the required language. The EU AI Act's requirements for transparency and documentation of AI system limitations are relevant for regulated deployments. 2. GOVERNANCE EXPOSURE: Low to Medium depending on deployment context. For general-purpose developer tools, this is a standard technical limitation disclosure. For consumer-facing regulated applications in non-primary language markets, the reliability limitation creates deployment risk that customers must independently evaluate. 3. JURISDICTION FLAGS: EU deployments in member states with non-primary supported languages (such as Dutch, Polish, or Swedish) should assess whether model output reliability is sufficient for consumer-facing regulatory communications. Deployments in markets where local language accuracy is a legal requirement face heightened exposure. 4. CONTRACT AND VENDOR IMPLICATIONS: Enterprise customers deploying in multilingual contexts should document which languages their application relies on and assess whether those languages fall within the 10 primary or 13 lower-resource categories. SLA-level reliability commitments for specific languages should be sought from Cohere's customer agreement rather than inferred from this model card. 5. COMPLIANCE CONSIDERATIONS: Compliance teams should assess whether their specific language requirements are met by the model's disclosed performance profile. For regulated consumer communications in non-primary languages, independent evaluation or supplemental translation review mechanisms may be warranted.
Full institutional analysis
Regulatory citations, enforcement risk, and due diligence action items.
Monitor: same-day alerts on the platforms you choose. Analyst: full institutional analysis.
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.
This disclosure establishes performance boundaries that are relevant for enterprise deployments serving multilingual populations, and creates due diligence obligations for customers deploying the models in contexts where output reliability in non-supported languages could affect user safety or regulatory compliance.
The document discloses that Command R model outputs in languages outside the 23 supported languages are described as unlikely to be reliable. For the 13 lower-resource languages within the 23, the document notes they are less rigorously evaluated than the 10 primary languages.
No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Cohere.