Cohere Command A+ Now Available in Microsoft Foundry
Summary
Cohere Command A+, Cohere's latest high-performance open-source mixture-of-experts (MoE) model, is now available in Microsoft Foundry as a Managed Compute offer. This model, released under an Apache 2.0 license, is designed for enterprise-grade agentic AI workloads, emphasizing efficiency and scalability. It integrates reasoning, multimodal understanding, multilingual capabilities, retrieval, and tool use, supporting a 128K context window and 48 languages. Command A+ is optimized for applications like Retrieval-Augmented Generation (RAG), agentic workflows, and enterprise search. Its sparse MoE architecture features 218B total parameters with only 25B active, enabling deployment on a single NVIDIA Blackwell GPU or two NVIDIA H100 GPUs. Cohere reports up to 63% higher output tokens per second and improved performance across enterprise benchmarks.
Key takeaway
For MLOps Engineers operationalizing open-source models, Cohere Command A+'s availability in Microsoft Foundry offers a streamlined path. You can deploy this efficient MoE model on dedicated managed infrastructure, reducing complex serving stack management. This integration allows you to scale AI workloads with greater control over performance and optimize throughput, while standardizing deployment and integrating governance into production systems. Consider this for your enterprise agentic AI applications requiring high efficiency and operational simplicity.
Key insights
Command A+ offers enterprise-grade agentic AI with high efficiency and managed deployment flexibility via Microsoft Foundry.
Principles
- MoE architecture enhances efficiency.
- Managed compute simplifies open-source model deployment.
- Unified platforms streamline AI operations.
In practice
- Deploy agentic AI workflows.
- Enhance RAG systems.
- Automate business processes.
Topics
- Cohere Command A+
- Microsoft Foundry
- Agentic AI
- Mixture-of-Experts
- Managed Compute
- Enterprise AI
Best for: AI Architect, NLP Engineer, CTO, AI Engineer, Machine Learning Engineer, MLOps Engineer
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by Microsoft Foundry Blog articles.