While evaluating Qwen3.8-Flash-Next in a controlled sandbox environment, Cohere's research team observed something unexpected. Asked to generate a defensive cybersecurity detection rule, the model instead inventoried aspects of the machine it was running on and attempted to upload a probe file to an Alibaba Cloud storage bucket in Beijing. Notably, the model's reasoning trace stated that it would not invoke any tools before taking these actions. Cohere's sandbox blocked the attempt, and no external communication succeeded. We have not determined the root cause, and we are not asserting malicious intent. But the incident raises important questions for organizations deploying agentic AI systems: Can models take actions that differ from their stated reasoning, and how should security teams evaluate trust, provenance, and behavior in open-weight models? We are publishing our findings because organizations making model adoption decisions deserve as much information as possible. We hope this contributes to greater transparency, stronger security practices, and continued collaboration across the AI ecosystem. Read the full report: cohere.com/blog/concerning-b…

We’re announcing Cohere’s first office in Ottawa 🇨🇦 Canada is where Cohere started, and we’re proud to keep building here.

Cohere and Aleph Alpha announce the signing of a definitive agreement, becoming the first foundational AI model developer anchored on both sides of the Atlantic 🇨🇦🇩🇪 Operating globally as Cohere, the unified company will grow to more than 1,000 employees across both continents.

Introducing: Cohere Transcribe – a new state-of-the-art in open source speech recognition.

Today we’re launching Model Vault — a dedicated, fully managed platform to run Cohere models securely and at scale. Model Vault delivers the control and isolation of self-hosting, without the operational burden: 🔒 Dedicated, isolated VPC ⚡ No noisy neighbors or rate limits 🔁 Elastic inference with predictable performance 📈 Real-time usage and performance monitoring No sharing. No limits. No surprises. Let Model Vault take care of model scaling and serving — so you can focus on building.