North Mini Code
MoEToolsThinkingCohere · 30B (3B active) · Mixture of Experts
Open agentic coding MoE with 3B active — built for software engineering and terminal tasks
0
0
2026-06
256K tokens
Use cases
Your hardware
16 GB Reference Desktop
Quantization options
What is quantization?| Quant | Bits | VRAM | Disk | Quality | Status |
|---|---|---|---|---|---|
| Q2_K | 2 | 10.1 GB | 9.2 GB | low | Decent · ~20 t/s |
| Q3_K_M | 3 | 13.9 GB | 12.8 GB | moderate | Barely runs · ~15 t/s |
| Q4_K_M | 4 | 15.9 GB | 14.7 GB | good | Barely runs · ~13 t/s |
| Q5_K_M | 5 | 19.7 GB | 18.3 GB | good | Barely runs · ~5 t/s |
| Q6_K | 6 | 23.6 GB | 22 GB | excellent | Barely runs · ~3 t/s |
| Q8_0 | 8 | 31.2 GB | 29.3 GB | excellent | Barely runs · ~2 t/s |
| F16 | 16 | 62 GB | 58.7 GB | lossless | Too heavy · ~1 t/s |
About this model
North Mini Code (north mini code) is a 30B sparse mixture-of-experts (MoE) model activating 3B active per token from Cohere in the North family, designed primarily for code, reasoning workflows. Native reasoning/thinking tokens expand output length per query, so reserve extra KV cache headroom. Supports structured tool calling; use Q4_K_M or higher when strict JSON schema adherence matters. Supports up to 256K context tokens; start at 4K–8K with Q8_0 KV cache on tight GPUs.
Hardware planning baselines start at 16.8 GB minimum memory and 27.9 GB recommended memory. At Q4_K_M, weights plus baseline runtime overhead use roughly 15.9 GB of RAM/VRAM and 14.7 GB of disk space. Because MoE routing evaluates only a subset of experts on each generated token, North Mini Code delivers faster decoding throughput per gigabyte of memory bandwidth than a dense checkpoint of identical total size—provided the full 15.9 GB checkpoint stays resident without SSD paging.
Deploy North Mini Code locally via Ollama (`north-mini-code-1.0`), Ollama with Flash Attention enabled. Under Apache License 2.0, review official weight cards before commercial redistribution.
Highlights
- Local Fit Baseline16.8 GB minimum and 27.9 GB recommended RAM/VRAM.
- Primary QuantQ4_K_M targets 15.9 GB memory and 14.7 GB disk.
- Context & Capabilities256K max context · Sparse MoE · Tool calling · Reasoning.
System requirements by device
Memory planning targets 16.8 GB minimum RAM/VRAM and 27.9 GB recommended memory. The baseline quant is Q4_K_M, estimated at 15.9 GB memory and 14.7 GB disk. Actual runtime usage shifts with context length, KV cache precision, and OS background overhead.
| Device | Notes |
|---|---|
| Mac (Apple silicon) | Use unified memory headroom (~75% physical RAM max) as the primary fit signal. Start with Q4_K_M and close heavy browser tabs first. |
| Windows / Linux GPU | Keep weights plus KV cache inside dedicated VRAM so Windows WDDM does not spill layers over PCIe into Shared GPU Memory. |
| Phones & Compact Edge Boards | Not practical on phones (16.8 GB min RAM exceeds Android LMK and iOS Jetsam limits). Pick a 2B–4B model on mobile. |
Start with a short session and leave memory for the operating system. A file fitting on disk does not guarantee that it will load in memory.
Estimated memory by quantization
These are rough estimates, not tested minimum requirements. Check the actual file and runtime documentation; context, offloading and operating-system overhead can change the memory needed.
| Quant Tier | Estimated memory | Disk Size | Quality |
|---|---|---|---|
| Q2_K | 10.1 GB | 9.2 GB | low |
| Q3_K_M | 13.9 GB | 12.8 GB | moderate |
| Q4_K_M | 15.9 GB | 14.7 GB | good |
| Q5_K_M | 19.7 GB | 18.3 GB | good |
| Q6_K | 23.6 GB | 22.0 GB | excellent |
| Q8_0 | 31.2 GB | 29.3 GB | excellent |
| F16 | 62.0 GB | 58.7 GB | lossless |
How to run North Mini Code on your device
Start with a short prompt and the smallest quant that still gives the quality you need. A clean first run is more useful than loading a larger file with no memory headroom.
Run North Mini Code on Mac
- Install Ollama or LM Studio on macOS Apple Silicon.
- Pull Q4_K_M (15.9 GB) first and confirm it stays below ~75% of your Mac's unified memory.
- Enable Flash Attention and cap initial context at 8K until memory pressure stays below 27.9 GB.
- Consult the Mac guide for Activity Monitor tuning and Metal offload tips.
Run North Mini Code on Windows
- Install LM Studio for Windows for visual VRAM inspection, or Ollama when you want a localhost API endpoint.
- Load Q4_K_M (~15.9 GB) and verify Dedicated GPU Memory stays within VRAM without spilling into Shared GPU Memory.
- Reserve at least 16.8 GB memory minimum and target 27.9 GB for comfortable context growth.
- Enable Q8_0 KV cache or drop one quant tier if generation slows down over PCIe.
North Mini Code vs alternatives
| Model | Params | Context | Min RAM | License |
|---|---|---|---|---|
| North Mini Code | 30B (3B active) | 256K | 16.8 GB | Apache 2.0 |
| Qwen 3 30B-A3B | 30B (3.3B active) | 128K | 16.8 GB | Apache 2.0 |
| Nemotron 3 Nano 30B | 30B (3B active) | 1M | 16.8 GB | NVIDIA Open |
| Granite 4.1 30B | 30B | 128K | 16.8 GB | Apache 2.0 |
| Qwen 3 Coder 30B-A3B | 30B (3B active) | 256K | 16.8 GB | Apache 2.0 |
| Muse Glimmer 30B | 30B | 128K | 16.8 GB | Apache 2.0 |
Use this table to choose the smallest nearby model that still matches the job. A related model with lower RAM can be the better local choice when the current page fits only at the edge of available memory.
Frequently asked questions
- What is North Mini Code?
- North Mini Code is a 30B sparse mixture-of-experts model (3B active active per token) from Cohere, supporting up to 256K tokens with 16.8 GB minimum RAM and 27.9 GB recommended RAM.
- How do I download and install north mini code?
- Download a GGUF build for North Mini Code, start with Q4_K_M, and confirm the model fits locally before increasing context or quant quality.
- How much VRAM or RAM does North Mini Code require?
- At Q4_K_M, North Mini Code occupies roughly 15.9 GB of active memory and 14.7 GB on disk. Add 2–6 GB of spare headroom for KV cache buffers, runtime scratch space, and OS apps (27.9 GB recommended).
- What is the best way to run North Mini Code locally?
- Use `ollama run north-mini-code-1.0` for a quick CLI/API setup, or load Q4_K_M in LM Studio or llama.cpp with Flash Attention enabled when you want explicit KV cache and layer control.
- Can North Mini Code run on Android or iPhone?
- North Mini Code requires 16.8 GB minimum memory, placing it outside practical 8 GB smartphone limits. Run it on a 16 GB+ Mac or desktop GPU instead, and pick a 2B–4B checkpoint for phones.
- How does North Mini Code compare to Qwen 3 30B-A3B?
- Qwen 3 30B-A3B (30B, 16.8 GB min RAM) sits closest in size. Compare active parameters, context ceiling (256K vs 128K), and license terms (Apache 2.0) before switching.