Model Detail

Qwen3.6-35B-A3B-4bit-DWQ is a large language model with 35B parameters released by mlx-community. The model is registered under the text-generation pipeline tag on Hugging Face.

Qwen3.6-35B-A3B-4bit-DWQ ships with 35B parameters. Total weight footprint is approximately 34.7 GB, which is the relevant figure when planning local-inference VRAM.

Qwen3.6-35B-A3B-4bit-DWQ is best fit for general-purpose chat and instruction-following workloads. Treat this as a starting matrix rather than a benchmark verdict — the right deployment usually depends on the specific evaluation suite that mirrors your workload.