Cohere just dropped its first open-source agentic coding model, and the architecture tells you everything about where the enterprise AI race is heading. North Mini Code 1.0, a 30 billion parameter Mixture-of-Experts model, launched on June 9 under the Apache 2.0 license, making it freely available on Hugging Face for anyone who wants to run a capable coding assistant without phoning home to someone else’s cloud.
The model scored 27.6 on the Artificial Analysis Intelligence Index, a benchmark that attempts to standardize how we compare AI models across capabilities.
The MoE trick that makes this work
North Mini Code routes each query to a small subset of specialized “expert” networks within the larger model, with 30 billion parameters total but only 3 billion active at any given time, keeping inference costs dramatically lower than a dense 30B model would require.
The model supports a context length of 256K tokens and can generate outputs up to 64K tokens. To put those numbers in perspective, 256K tokens is roughly equivalent to feeding the model an entire mid-sized codebase and asking it to understand the relationships between files, functions, and dependencies. The 64K output ceiling means it can generate substantial blocks of code in a single pass rather than requiring developers to chain together multiple shorter responses.
