Back to registry
CUDA24GB VRAM

arnav080/qwen3.6-27b-dual-gpu-mtp-above_spec

Optimized Qwen 3.6 27B dual-GPU recipe with native MTP speculative decoding & 100k context.

Configuration Specifications

Base Modelhuggingface:Qwen/Qwen3.6-27B-Instruct
Enginellama.cpp
QuantizationQ4_K_M
Platformcuda
VRAM Required24GB minimum

Telemetry Benchmarks

No runs recorded yet

Be the first to benchmark this recipe! Run the CLI command from your terminal:

bloc run arnav080/qwen3.6-27b-dual-gpu-mtp-above_spec
qwen3.6-27b-dual-gpu-mtp-above_spec.yaml
Loading workspace editor...