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 Model | huggingface:Qwen/Qwen3.6-27B-Instruct |
| Engine | llama.cpp |
| Quantization | Q4_K_M |
| Platform | cuda |
| VRAM Required | 24GB 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...