Is anyone getting real coding work done with Qwen3.6-35B-A3B-UD-Q4_K_M on a 32GB Mac in opencode, claude code or similar?

A user is attempting to perform real coding tasks with Qwen3.6-35B on a 32GB M2 Macbook Pro, encountering memory exhaustion and context window management issues. Despite the model identifying the essence of a bug, it struggles with implementation as critical information is lost during context compaction.