Is it possible to disable thinking via API request to Qwen3.6-35B-A3B?

[zaskara]

I use the OpenAI v1/responses. I tried to add: reasoning: { "effort": "none" }, but it didn't turn off thinking.

[deepsweet]

Hi. Have you tried "chat_template_kwargs": {"enable_thinking": false}?

[zaskara]

Thank you, I added this and it works:
chat_template_kwargs: { enable_thinking: false, preserve_thinking: false }