just curious, why the lack of <thinking> response?
#4
by
YellowjacketGames
- opened
No complaints, just interested in why. Needless tokens when a coding model is just trying to get stuff done?
Probably a better question for https://huggingface.co/Qwen/Qwen3-Coder-Next since this repo is a quantization of the upload by the Qwen team, meaning they were the ones that made the decision to omit reasoning.
True, for some reason I was reading this as a feature of their quant, not a feature of the underlying model.
It doesn't actually make sense for an agentic coding model. Since the agentic framework itself -- the backlog of messages in the tool -- provides the history of what was going on and what was tried and what is currently being tried, it effectively already IS a thought process.