LLM Risk Profile

Cohere (Command R+)

Toronto-based AI company founded 2019 by ex-Google researchers. Builds secure, enterprise-grade LLMs for regulated industries like finance/healthcare. Focus: customizable, on-prem deployable models enhancing productivity via agentic AI like North platform.Cohere About, Wikipedia

Models: Command R+ & Embed
Flagship Models
Command A (latest flagship, excels in agents/RAG/tools, 256K context), Command R+ (optimized for complex RAG/multi-step agents, 128K context)
Enterprise Adoption
~$240M ARR 2025 (beat target), 70% margins, multi-cloud. Strong in regulated sectors.CNBC
Pricing
Mid-market: Command R+ $2.5/1M input, $10/1M output (cheaper than Anthropic). Free trial 1K calls.Pricing
Overview

About Cohere

Toronto-based AI company founded 2019 by ex-Google researchers. Builds secure, enterprise-grade LLMs for regulated industries like finance/healthcare. Focus: customizable, on-prem deployable models enhancing productivity via agentic AI like North platform.Cohere About, Wikipedia

Agentic AI

Agentic Capabilities

Command R+ supports multi-step tool use/agents, RAG w/ citations/grounding, improved tool decision-making (Aug 2024 update). Enables reasoning/dynamic actions over tools for production workflows.Cohere Docs, Models

Deploying Cohere in Production?

Get comprehensive coverage for your entire AI technology stack.

Insure Your AI Strategy →
Incident History

Known Incidents & Failures

Feb 2025: Sued by Condé Nast/McClatchy for copyright infringement (reproduced full articles, hallucinated fake news).TVNewsCheck Aug 2024: Jailbreak prompt bypasses filters.GitHub No data breaches/leaks found.

Risk Analysis

Comprehensive Risk Profile

Detailed breakdown of every risk category for enterprises deploying Cohere models in agentic AI workflows.

⚠️ Hallucination & Confabulation Risks

Can generate fabricated content attributed to sources (e.g., fake Guardian article in lawsuit). RAG mitigates but not eliminates; acknowledged business risk.Lawsuit, Cohere Blog

🛡️ Data Privacy & Leakage Risks

SOC2 Type II, GDPR DPA w/ EU-US TIA, opt-out training, VPC/on-prem deploys, encryption. No incidents; general API key/prompt injection risks (mitigated e.g., 86% on Aya).Security, Trust Center

🔒 Bias & Discrimination Risks

Training data biases (race/gender/etc.); BOLD dataset eval. Potential discriminatory outputs in enterprise use.Model Card, SaferAI

⚖️ Security & Jailbreak Vulnerabilities

Jailbreaks via persona prompts (2024); prompt injection reducible. Safety modes/post-deploy backdoors noted.GitHub, Hirundo

🎭 Unauthorized Autonomous Action Risks

Agentic tool chains vulnerable post-jailbreak, potentially executing unintended API calls despite decline training.Agent Jailbreak

© Model Drift & Reliability Concerns

Regular updates (e.g., R+ 08-2024: +50% throughput, better instructions). Leaderboard selection bias noted in Cohere paper.Models

📉 Regulatory & Compliance Risks

GDPR/CCPA compliant; ongoing copyright suit. EU AI Act high-risk for agents.Trust Center

📜 IP & Copyright Infringement Risks

2025 lawsuit: trained on 4K+ unauthorized works, outputs full paywalled articles.TVNewsCheck

🔐 Deepfake & Misuse Potential

Text-focused; low native deepfake risk. Jailbreak-enabled harmful text gen possible.SaferAI

Coverage Needs

Insurance Implications

E&O for hallucinations/agent errors; cyber for breaches/injection; D&O for fines/litigation. Emerging AI policies; check exclusions.eSpecialty, Resultsense

Who Uses Cohere

Notable Enterprise Customers

RBC, LG, McKinsey, Salesforce, Nvidia, Oracle, Notion, Google.CBInsights, Video

Don't Let AI Risk Become Business Risk

Don't let model failures become business failures. Get covered today.

Protect Your AI Deployment →
Related Risks

Risk Categories for Cohere

Get Covered

Recommended Insurers

Explore More

Other LLM Providers