AI DevOps + OpenCode: Make Coding Free with Zen Models - NO Claude, OpenAI, Google

[robi]

OpenCode + AI DevOps: Free Coding with Zen Models

How to use OpenCode's free Zen models for coding — no GPU, no Ollama, no API key needed.

Following up on the Ollama thread — there's an even simpler way to code for free: OpenCode's built-in Zen models.

No local models to pull. No GPU required. No API key to manage. Just install AiDevOps with OpenCode and start coding.

What is OpenCode Zen?

OpenCode Zen is a curated set of models served by the OpenCode team. Some are free:

Model	ID	Best For	Cost
Big Pickle	big-pickle	General coding, stealth model	Free
MiMo V2 Pro Free	mimo-v2-pro-free	Xiaomi's coding model	Free
MiniMax M2.5 Free	minimax-m2.5-free	Strong coding, long context	Free
Nemotron 3 Super Free	nemotron-3-super-free	NVIDIA's model	Free

These are the same models that power paid coding assistants — just served for free through OpenCode.

How to Choose a Free Model

1. Install AI DevOps: npm install -g aidevops && aidevops update
2. The TUI opens for you: just run opencode after install.
3. Type /models to see available models
4. Select a free one (arrow keys up/down - look for "Free" in the name)

Or from the CLI:

opencode run --model opencode/big-pickle "explain this code"

How It Works with aidevops

When you're running aidevops (the AI DevOps framework by our own @marcusquinn), OpenCode is the default coding agent. The free Zen models work out of the box, just select them, it remembers.

The aidevops pulse supervisor dispatches coding tasks to OpenCode, which uses the free Zen models. No configuration needed.

Model Selection Guide

For everyday coding: Big Pickle

• Fast, free, handles most tasks
• Good for file edits, function implementations, code reviews

For complex reasoning: MiniMax M2.5 Free

• Stronger reasoning, longer context
• Good for architecture decisions, debugging complex issues

For long files: MiMo V2 Pro Free

• Handles large codebases well
• Good for refactoring, multi-file changes
• Tries hard, I like the way it works, a fav.

For quick tasks: Nemotron 3 Super Free

• Fastest response time
• Good for simple edits, formatting, documentation

Switching Models

In the OpenCode TUI:

/models → select model → done

From CLI:

opencode run --model opencode/minimax-m2.5-free "refactor this function"

Privacy Note

Free models may use your data for training during the free period. If you're working on sensitive code, use a privacy model or run locally with Ollama / LM Studio.

Cost Comparison

Setup	Cost	Setup Time
OpenCode Zen (free models)	$0	2 minutes
Ollama + local models	$0 + GPU	30 minutes
Claude Code + API	$3-15/M tokens	5 minutes
GitHub Copilot	$10/month	5 minutes

OpenCode Zen is the fastest path to free coding. No hardware, no subscriptions, no limits (within reason).

Links

• OpenCode: https://opencode.ai
• OpenCode Zen docs: https://opencode.ai/docs/zen/
• AI DevOps: https://aidevops.sh
• Free models list: https://opencode.ai/docs/zen/#pricing

---

Tested on macOS 26 with OpenCode 1.2.27 and aidevops v3.1.28.