Comment by selcuka

Would you call Qwen 3.6 35B "similar size" with ~50% more parameters (both active and total) though? I believe what GP meant that if there was a Qwen 3.8 24B-A2B, it would roughly be as intelligent as this one.