A local coding agent struggled even with detailed plans
A local coding setup using Qwen 3.6 27B UD_4 did not work well for coding help. The machine used an RTX 5090, a 131K context, no , and a model quantized by Unsloth. Cline inside VS Code was chosen because it seemed close to Claude Code.
Fable created detailed, step-by-step plans for a basic Python app, written clearly enough for a junior engineer to follow, and those plans were placed in the repo. Qwen still produced code with many mistakes and generated with basic syntax errors. It could not reliably follow the detailed plan, even when the task was mostly to move through the steps.
The expected level was not Claude Opus or , but the setup still failed at basic coding-agent work through the tool harness and terminal calls.
Key points
- The setup used Qwen 3.6 27B UD_4 on an RTX 5090.
- It ran with a 131K context and no .
- Cline in VS Code was used as the coding-agent tool.
- Detailed Fable plans for a basic Python app did not prevent code mistakes or broken .
- Local models may reduce direct API costs, but only if they can complete agent tasks reliably.