OpenClaw
Updated: February 12, 2026
This guide walks through connecting SiliconFlow to OpenClaw via "OpenAI Compatible Mode" to access 100+ high-performance large models on the SiliconFlow platform.
Step 1: Prepare SiliconFlow API
Before configuration, make sure you have the following:
-
SiliconFlow API Key
- A string starting with
sk-.
- A string starting with
-
Base URL
- URL:
https://api.siliconflow.cn/v1 - Important: Make sure to include the trailing
/v1
- URL:
-
Model Name
- Copy the full model name from the SiliconFlow Models page, e.g.,
deepseek-ai/DeepSeek-V3
- Copy the full model name from the SiliconFlow Models page, e.g.,
Step 2: Open the openclaw.json config file
(Using macOS as an example, the default file path is: ~/.openclaw/openclaw.json)
- Copy and merge the following into your config file (replace your API Key and desired model id. Note: if you have already configured custom models, start merging from the third line.
"models": {
"providers": {
"siliconflow": {
"baseUrl": "https://api.siliconflow.cn/v1",
"apiKey": "YOUR_SILICONFLOW_API_KEY",
"api": "openai-completions",
"models": [
{
"id": "zai-org/GLM-4.6",
"name": "GLM-4.6",
"reasoning": false,
"input": [
"text"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 200000,
"maxTokens": 8192
}
]
},- Merge the following into the
agentssection of openclaw.json (as shown in the image).
Note: The model id in
siliconflow/zai-org/GLM-4.6must match the model id you configured in the previous step. It is recommended to add the model tofallbacksfirst to avoid errors caused by incorrect configuration.
"siliconflow/zai-org/GLM-4.6",
Step 3: Verify Configured Models
Run openclaw models list in the terminal to see your configured SiliconFlow models.

Step 4: Switch Models
In the web interface or your configured chat app, type: /model siliconflow/zai-org/GLM-4.6 to temporarily switch to the SiliconFlow model you just added.
