Qwen3.6-27B-NVFP4 - images

Model: Abiray-Qwen3.6-27B-NVFP4.gguf Specs: - Legion 7i Gen10 - NVIDIA GeForce RTX™ 5090 - Intel® Core™ Ultra 9 275HX × 24 - RAM 32.0 GiB llamacpp settings: ./build/bin/llama-server \ -m ~/.lmstudio/models/lmstudio-community/Qwen3.6-27B-GGUF/Abiray-Qwen3.6-27B-NVFP4.gguf \ -ngl 99 \ -c 131072 \ -t 16 \ -b 4096 \ -ub 2048 \ --cache-type-k q8_0 \ --cache-type-v q8_0 \ -fa 1 \ --defrag-thold 0.1 \ --temp 0.6 \ --top-p 0.95 \ --top-k 20 \ --min-p 0.0 \ --presence-penalty 0.0 \ --repeat-penalty 1.0 \

More from LocalLlama

Hybrid on-device inference on Android: llama.cpp + LiteRT + NPU/GPU routing

LocalLlama·4

MiniMax M2.7 AWQ-4bit on 2x Spark vs 2x RTX 6000 96GB - performance and energy efficiency

LocalLlama·3

Qwen3.6-27B at 72 tok/s on RTX 3090 on Windows using native vLLM (no WSL, no Docker), portable launcher and installer

LocalLlama·5

A Dark-Money Campaign Is Paying Influencers to Frame Chinese AI as a Threat

LocalLlama·5

"LLM is created so engineer don't have to write a report", anyway found out ONLYOFFICE can connect to OpenAI compatible, using Qwen 3.6 to do elaboration.

LocalLlama·5

Blog: AI evals are becoming the new compute bottleneck

LocalLlama·6