Comment by hadlock

By their own benchmarks it is about 10% lower scoring than Qwen 3.6 35b-a3b, but I've added it to my list. Always looking for MoE to compare to it so we can squeeze more out of our local LLM system.