Command R (08-2024) is a 32-billion parameter generative large language model developed by Cohere and Cohere Labs as part of the Command R family of foundation models. It was released as a research-oriented refresh of the Command series in August 2024, with an accompanying official announcement and changelog describing the updated lineup. The model is designed for robust performance in retrieval-augmented generation, multilingual comprehension, reasoning, and advanced tool interactions, positioning it for both enterprise and research workloads that need a balance of capability and efficiency.

Weights for Command R (08-2024) are openly distributed under a CC-BY-NC 4.0 license, with the full model hosted on Hugging Face under CohereForAI and quantized GGUF builds available for local inference, including an Ollama distribution. This open availability makes the model well suited for teams that want to run it on their own infrastructure or experiment with different precision tradeoffs. Its focus on RAG pipelines, reasoning, and tool use aligns it with practical applications such as grounded question answering, multi-step assistants, and workflows that need reliable integration with external systems.