North Mini Code Is Now Available in Puter.js

Puter.js now supports North Mini Code, Cohere Labs' open-weight coding model built for agentic software engineering. It is free to use with no API keys required.

What is North Mini Code?

North Mini Code is an open-weight agentic coding model from Cohere Labs, released under the Apache 2.0 license. It uses a sparse Mixture-of-Experts architecture with 30B total parameters and 3B active per token, drawing from 128 experts with 8 activating on each token. This sparse design lets it run on a single H100 GPU at FP8 precision.

Key highlights:

• 256K context window: Work across multi-file, repository-scale codebases and long agent traces in a single prompt, with up to 64K output tokens per response
• Sparse MoE architecture: 30B total parameters with only 3B active per token, keeping inference fast and inexpensive
• Agentic by design: Supports interleaved thinking and native tool use for sub-agent orchestration, multi-file analysis, and code review
• Coding-first focus: Optimized for code generation, agentic software engineering workflows, and real terminal tasks
• Open weights: Published on Hugging Face under the Apache 2.0 license

Performance

North Mini Code scores 33.4 on the Artificial Analysis Coding Index, outperforming open-source models up to 4x its size. It posts competitive results on SWE-Bench Verified, SWE-Bench Pro, and Terminal-Bench 2.0.

Compared to Devstral Small 2, it delivers 2.8x higher output throughput and 30% lower inter-token latency, which makes it a practical choice for coding agents and automated engineering pipelines where speed and cost matter.

Examples

Code generation

puter.ai.chat(
  "Write a Python function that returns the nth Fibonacci number using memoization, with tests.",
  { model: 'cohere/north-mini-code:free' }
);

Repository-scale code review

puter.ai.chat([
  { role: "user", content: largeCodebase + "\n\nReview this codebase for bugs, dead code, and opportunities for simplification. Group findings by file." }
], { model: 'cohere/north-mini-code:free' });

Agentic refactoring

puter.ai.chat(
  "Refactor this Express route to use async/await, add input validation, and propagate errors to the error-handling middleware. Explain each change.",
  { model: 'cohere/north-mini-code:free' }
);

Streaming

const response = await puter.ai.chat(
  "Design a token-bucket rate limiter in Go with unit tests covering burst and steady-state traffic.",
  { model: 'cohere/north-mini-code:free', stream: true }
);

for await (const part of response) {
  puter.print(part?.text);
}

Get Started Now

Just add one library to your project:

// npm install @heyputer/puter.js
import { puter } from '@heyputer/puter.js';

Or add one script tag to your HTML:

<script src="https://js.puter.com/v2/"></script>

No API keys and no infrastructure setup. Start building with North Mini Code immediately.

Learn more:

• Free, Unlimited Cohere API
• Puter.js AI Documentation