Cohere’s 30B Open-Weight North Mini Code Targets Agentic Coding Workflows
Cohere’s North Mini Code is a 30B-parameter open-weight mixture-of-experts model with 3B active that builds on Command A+ and is aimed at agentic coding workflows rather than one-shot code prompts. The model was described as being designed for tasks such as using a terminal, inspecting an environment and working through software-repository fixes, including evaluations on Terminal-Bench and SWE-Bench.
Released earlier this week, North Mini Code already has MLX support and can run locally through llama.cpp, Ollama or vLLM, with quantized versions ranging from 9GB to full BF16. Cohere also stressed that users can keep using North Mini Code and Command A+ after access to Claude Fable 5 was suspended under a US government directive.
From the sources (6 posts)
@cohereWe released North Mini Code, our first open-source coding model, just two days ago. We’re already seeing developers using it to build amazing things ⬇️
@cohereDay zero MLX support from @Prince_Canuma
@cohereOur community moves fast. Unsloth GGUFs are already up. Run North Mini Code locally with a llama.cpp setup, Ollama, or vLLM, with quants from 9GB to full BF16
@rasbtCool new open-weight model by Cohere: a new lightweight 30B open-weight model for agentic coding tasks. This one builds on Command A+ using the parallel transformer design. Interestingly, even though it's almost half as big, it almost doub
@cohereYou can continue to use Command A+ and North Mini Code whether we want you to or not
@quixiaiRT @cohere: You can continue to use Command A+ and North Mini Code whether we want you to or not