Cohere Introduces Command A Reasoning for Enterprise Workflows

Cohere has introduced Command A Reasoning, a new language model tailored for complex reasoning tasks in enterprise settings. Announced on their website, the model excels in agent workflows, large document processing, and end-to-end systems, outperforming other private models like gpt-oss-120b and DeepSeek-R1 0528.

Command A Reasoning is designed to run efficiently on a single H100 or A100 GPU with a context window of 128,000 tokens, and can scale up to 256,000 tokens across multiple GPUs. This flexibility allows enterprises to manage resource use and costs effectively by setting a token budget, enabling parallel workloads without switching models.

The model is available as a research release with open weights under a CC-BY-NC-4.0 license for non-commercial use, while commercial deployments require contacting Cohere's sales team. It also incorporates a safety framework to balance strict abuse prevention with minimizing false rejections of legitimate requests.

Command A Reasoning is now accessible on the Cohere platform and on Hugging Face, providing enterprises with a powerful tool for enhancing their AI-driven research and workflows.