Best local model supporting claude code? Rtx3060

Posted by CatSweaty4883@reddit | LocalLLaMA | View on Reddit | 2 comments

Hello all, I’ve been using Qwen 3.5 9B Q4 262k ctx using Llama cpp for claude code for a while now, is there any model which better complements agentic coding setup locally? Or is there a better harness (than Claude Code)?

System RAM: 16GB
VRAM: 12GB