North Mini Code Community Tops 150K Downloads

// 1h agoPRODUCT UPDATE

North Mini Code Community Tops 150K Downloads

Cohere says its open-source North Mini Code model has surpassed 150,000 downloads as developers continue experimenting with agentic coding workflows. With OpenCode’s free tier ending, developers can access it through Model Vault, OpenRouter, or local deployment.

// ANALYSIS

North Mini Code’s momentum shows that efficient, open-weight coding models can attract serious developer interest when they offer flexible deployment and low operating costs.

–Its 30B-parameter MoE architecture activates roughly 3B parameters per token, improving efficiency for coding workloads
–Apache 2.0 licensing supports local, self-hosted, and commercial experimentation
–OpenRouter and Model Vault preserve hosted access after OpenCode’s free-tier change
–Local deployment gives developers more control over privacy, latency, and vendor dependence
–The 150K-download milestone is promising, though sustained adoption will depend on real-world coding-agent performance

// TAGS

north-mini-codellmopen-weightsai-codingcoding-agentopen-sourcelocal-firstinference

DISCOVERED

1h ago

2026-08-14

PUBLISHED

1h ago

2026-08-14

RELEVANCE

9/ 10

AUTHOR

cohere

// KEEP READING

More AI developer news from the feed

EXPLORE FULL FEED

UPDATE23m ago

Super.engineering adds GitHub PR reviews

super.engineering now lets developers review and respond to GitHub pull-request comments without leaving its native GPUI workspace. A configurable one-click action can trigger an automated review flow alongside existing agent sessions, diffs, checks, and merge actions.

MODEL25m ago

Qwen3.8-27B Brings Frontier Coding Local

Alibaba’s new 27B Qwen3.8 model is now available for local deployment, with early benchmarks showing it trading blows with or surpassing Claude Opus 4.6 on several coding tasks. The results could make frontier-level open-weight coding practical on consumer hardware.

MODEL27m ago

Qwen3.8-27B brings multimodal AI locally

Qwen released the open weights for Qwen3.8-27B, a 27B multimodal model built for coding and office workflows. It offers 262K native context, up to 1M tokens with YaRN, and an Apache 2.0 license.