Cohere has released North Mini Code 1.0, its first open-source agentic coding model, under the permissive Apache 2.0 license. The model has 30 billion total parameters but activates only 3 billion at inference time, suggesting a sparse architecture optimized for efficiency. It scores 33.4 on the Artificial Analysis Coding Index, positioned as competitive among models of comparable size, and is available on Hugging Face.
Cohere has introduced North Mini Code, a smaller, code-specialized variant of its North model family designed for developer use cases. The mini model prioritizes low latency and cost efficiency while retaining strong code completion, debugging, and explanation capabilities. This follows the industry trend of pairing flagship models with lightweight alternatives for high-frequency API usage in enterprise and individual developer contexts.
Microsoft announced MAI-Thinking-1, a 35B reasoning model available to select early partners, and MAI-Code-1-Flash, a 5B coding model rolling out to GitHub Copilot individual users in VS Code. Simon Willison highlights their relatively small parameter counts and Microsoft's claim that MAI-Thinking-1 was preferred to Sonnet 4.6 in internal blind evaluations. He also questions what Microsoft's clean and appropriately licensed training data claims mean in practice.