I've been using Qwen3.6 35B A3B, and with reasoning turned on, I'd say 2/3 (give or take) of the tokens for a response are thinking tokens. Which at 70+ tps locally, that isn't that awful. I run an 80k context across 4-10 "agents" for my solo TTRPG, where Qwen is the GM, each NPC at a location, the director, and the narrator.

Each turn is about 45-60 seconds to generate all of the various responses. The GM and director have reasoning on, and the NPCs/Location/Narrator do not.

It's a fairly good "engine" for that. I'm not sure how a denser Qwen would do here regarding speed.

Just to play devil’s advocate: you can’t compare Qwen to a (proprietary/closed source) hosted model and deduce that Qwen is overthinking, as Qwen gives you the full reasoning/thinking trace while all the proprietary models now give you only a summary “to prevent distillation”, making it hard to properly compare apples to apples here.

Qwen thinking is really good in Mandarin; and probably natively trained the most there.

Try a system prompt requiring it to think in Mandarin, while still delivering the response in the user’s language.

Disable thinking? I think many harnesses disable thinking on Qwen anyways because it interferes with tool calling.

Llamscpp provides reasoning budget and message. You can use the message to redirect it.

Once you get the agent and message consistent,itll keep moving.