The vLLM v0.23.0 release notes (source) include 'Cohere Mini Code' as a newly supported model, indicating Cohere has...

The vLLM v0.23.0 release notes (source) include 'Cohere Mini Code' as a newly supported model, indicating Cohere has released a code-specialized variant. This is a new product launch for Cohere, expanding its model portfolio beyond general-purpose and enterprise offerings. The release notes detail extensive improvements across many models, but the addition of Cohere Mini Code is highlighted in the 'New models' section. While Cohere's official website (source) emphasizes enterprise AI deployment and data sovereignty, it does not yet announce this specific model. Nonetheless, the inclusion in a major open-source inference engine like vLLM suggests the model is ready for production use and could drive incremental demand for Cohere's API services and compute inference infrastructure. The launch may also signal Cohere's focus on code generation and developer tools, potentially increasing competition with other code models like Code Llama and StarCoder. This could lead to higher adoption and compute requirements for inference workloads, benefiting AI infrastructure providers.