Skip to main content
Last Updated: March 24, 2026
Deploy from the SaladCloud Portal.

​
Overview

This recipe runs Qwen3.5-9B with the official SGLang runtime on a Salad GPU. The model is downloaded automatically on first startup, and the container exposes an OpenAI-compatible API for tools such as OpenClaw, OpenCode, and other compatible clients. This recipe is designed to be easy to use:
  • the model is already chosen for you
  • it is public by default, so you can test it immediately after deployment
  • it is preconfigured for Qwen reasoning and tool-calling
  • you can connect it to OpenClaw or any other OpenAI-compatible client

​
Quick Start

  1. Open the SaladCloud Portal.
  2. Open your Organization and click “Deploy Container Group”
  3. Pick Qwen3.5-9B (SGLang) from the recipe list.
  4. Enter a Container Group Name.
  5. Decide whether to enable Require Container Gateway Authentication:
    • Disabled: public access.
    • Enabled: requests must include your SaladCloud API key.
  6. Deploy and wait for the first startup to finish.
The model is downloaded from Hugging Face at startup, so it can take several minutes before the deployment becomes ready.
Once the container is ready, connect your client to https://<your-dns>.salad.cloud/v1/chat/completions.

​
Use With OpenClaw

If you want to connect this recipe to OpenClaw, follow this guide:

​
Current Defaults

The recipe currently defaults to:
  • Model source: Qwen/Qwen3.5-9B
  • Served model name: qwen3.5-9b
  • Host bind: ::
  • Runtime image: lmsysorg/sglang:latest-runtime
  • Context length: 131072
  • Tensor parallel size: 1
  • Memory fraction: 0.8
  • Attention backend: triton
  • Reasoning parser: qwen3
  • Tool call parser: qwen3_coder
  • Authentication: disabled by default
This follows the official Qwen SGLang launch pattern for Qwen/Qwen3.5-9B, with a triton attention backend selected for better compatibility on Salad GPU nodes.

​
Thinking Mode

Qwen thinking is enabled by default. If you want a direct response for a specific request, include this field in the JSON body: 
"chat_template_kwargs": {"enable_thinking": false}

​
Authentication

Require Container Gateway Authentication is available in the deployment form and is unchecked by default.
  • Disabled: anyone with the URL can call the API.
  • Enabled: every request must include the Salad-Api-Key header.
If you enable authentication, see Sending Requests for the header format.

​
Example Request

curl https://<your-dns>.salad.cloud/v1/chat/completions \
  -X POST \
  -H 'Content-Type: application/json' \
  -d '{
    "model": "qwen3.5-9b",
    "messages": [
      {"role": "system", "content": "You are a helpful assistant."},
      {"role": "user", "content": "Write a short explanation of mixture-of-experts models."}
    ],
    "temperature": 0.6,
    "top_p": 0.95,
    "top_k": 20,
    "max_tokens": 512
  }'
If you enabled authentication during deployment, add: 
-H 'Salad-Api-Key: <api-key>'

​
For Technical Users

If you want to change the model or runtime settings later, open the container group in the SaladCloud Portal and edit Advanced Configuration. Useful environment variables include:
  • SGLANG_MODEL_PATH
  • SGLANG_SERVED_MODEL_NAME
  • SGLANG_CONTEXT_LENGTH
  • SGLANG_MEM_FRACTION_STATIC
  • SGLANG_TP_SIZE
For more detail, see:

​
Source Code