How do I call opensource models

[AIExplorer25]

Hi Guys,
I want to call qwen3 model or other open source model instead of paid apis like gemini or anthropic models, my use case is the langgraph server is on a production server with only CPU and my qwen3 model is hosted on a different server with GPU. So I want to call the qwen3 model on different server and use it with existing code base, I suppose there should be example code for this somewhere, could anyone help me to find out that code snippet or is it something I need to write from scratch?

[ahmedzak1]

I'm using ollama for opensource models

model = ChatOllama(
model=model,
base_url="baseurl",
temperature=0)

agent = create_deep_agent(
model=model,
tools=[tavily_search, think_tool],
system_prompt=INSTRUCTIONS,
subagents=[research_sub_agent],
)

you can check https://docs.langchain.com/oss/python/integrations/providers/overview