Qwen3.6-35B-A3B MTP GGUF Models Uploaded
Unsloth Qwen3.6-35B-A3B UD XL models with grafted MTP layers released on HuggingFace in GGUF format. Benchmarks show modest 2.5-6% speedups on single RTX 5090 but up to 50% on multi-GPU setups like 2x 5070 Ti + 3090. Includes convert.py script; results vary by hardware.






