1 Install

Install Vs Code Insiders from https://code.visualstudio.com/insiders/ it's a beta version, unstable, seems great for testing agents, BUT, You have to be connected to the internet

LOOKS like it is deprecated

THIS IS THE JSON THAT WORKS

[
    {
        "name": "LM Studio Local",
        "vendor": "customendpoint",
        "apiKey": "${input:chat.lm.secret.-24f31728}",
        "apiType": "chat-completions",
        "models": [
            {
                "id": "qwen/qwen3.6-35b-a3b",
                "name": "qwen3.6-35b-a3b",
                "url": "http://127.0.0.1:1234/v1",
                "toolCalling": true,
                "vision": true,
                "maxInputTokens": 32768,
                "maxOutputTokens": 16000
            }
        ]
    }
]

This is the only model that is working


                            qwen3.6-35b-a3b

2

Select models from bottom of chat, click the gear icon then click the + add models button

3.

copy correct model name, for id then give in name for name

copy the correct info from LLM, make sure url container /v1 endpoint

copy the context length from llm model and put it in maxInputTokens

4

Select the model in options. Dope , your good to go, but, this will only work with internet,

Next lesson for pi