>it uses a lot of tokens to get work done.

That's irrelevant when you use $/task as the metric, which the OP does use.

It felt like a rocket compared to GLM 5.2 though. Are Chinese models generally token-heavy?

If I had the GPU size, hook it up to llama.cpp and setup the --reasoning-budget and reasoning-message; Most of that additional reasoning is a lot of garbage and you can redirect it to useful output.

That's how I handle the Qwen27B and 35B