[7900XT] Qwen3.6 27B for OpenCode

via

LocalLlama

I'm just looking for some advice on optimally setting up Qwen3.6 27B for OpenCode. The VRAM is a little bit scarce, but I ended up with this so far: llama-server --model models/Qwen3.6-27B-IQ4_XS.gguf \ --port 8080 \ --host 127.0.0.1 \ --top-p 0.95 \ --top-k 20 \ --min-p 0.0 \ --temperature 0.6 \ --flash-attn on \ --cache-type-k q8_0 \ --cache-type-v q8_0 \ --presence-penalty 0.0 \ --repeat-penalty 1.0 \ --ctx-size 65536 \ --chat-template-kwargs '{"preserve_thinking": true}' \ With this my VRAM

More from LocalLlama

daily ritual at this point…

LocalLlama·16

convert : add support for Nemotron Nano 3 Omni by danbev · Pull Request #22481 · ggml-org/llama.cpp

LocalLlama·7

Mistral Medium Is On The Way

LocalLlama·15

Something from Mistral (Vibe) tomorrow

LocalLlama·8

Deepseek Vision Coming

LocalLlama·10

First direct side by side MoE vs Dense comparison.

LocalLlama·11