zai-org

/

GLM-4.7-Flash

like 1.63k

Follow

Z.ai 11.2k

I need this feature as well. In the past, Qwen did it with just a single token. It was a nice feature but I don't know if this model allows to do this

Just like with GLM-4.7, you can view the operation methods in our GitHub.

For LmStudio add line:
{% set enable_thinking = false %}
to top of the 'Prompt Template'
I got screen from Glm-4.6V flash and Usloth gguf but it works for eveything the same: