How do I know if a model is "thinking switchable"?

[dlukt]

https://docs.continue.dev/customize/deep-dives/autocomplete
It says # qwen3 is a thinking-switchable model
How do I know it is? Using ollama right now

Also, what if I already have the 30b model installed and running, can I just use that instead of the 4b model?

[bdougie]

How do I know it is? Using ollama right now

I just searched Ollama desktop and confirmed it is. My suggestion is double checking the Ollama docs, but a simple query to an LLM will give you the answer below.

Also, what if I already have the 30b model installed and running, can I just use that instead of the 4b model?

Yes, you can absolutely use the 30B model instead of the 4B! Here's how:
yamlmodels:

  - name: Qwen3 30B without Thinking for Autocomplete  
    provider: ollama
    model: qwen3:30b  # or whatever your exact model name is
    roles:
      - autocomplete
    requestOptions:
      extraBodyProperties:
        think: false  # turning off the thinking for speed
        

The 30B model will likely give better autocomplete suggestions. The 4B model will be faster for autocomplete (which is usually preferred)