Support for Qwen 3.5 35B local model

[pkimtani]

Support for Qwen 3.5 35B model

Currently, when trying to run the docker-agent with a local model Qwen 3.5 35B, I keep getting jinja template error.
I verified my model setup (via LMStudio) that it works with other chat clients

[dgageot]

Hi @pkimtani could you share the image name of the model?

[pkimtani]

Hi @dgageot !
Did you mean this: qwen/qwen3.5-35b-a3b?

Additional link:

• https://lmstudio.ai/models/qwen/qwen3.5-35b-a3b

[dgageot]

@pkimtani I can't find a way to reproduce

I've tried with this agent:

providers:
  lm_studio:
    api_type: openai_chatcompletions
    base_url: http://192.168.1.123:1234/v1

agents:
  root:
    model: lm_studio/qwen/qwen3.5-35b-a3b
    instruction: You are a useful assistant
    toolsets:
      - type: filesystem
      - type: shell

Can you share the agent you're using and the type of questions that makes it fail?