Unable to Set Default Model llama3.1 in CrewBase – Model Not Found Error

CrewAI Community Support
,
1

I configured the default project when running crewai create and selected the default model as llama3.1. I’m trying to change it within CrewBase, but it doesn’t seem to take effect. I keep getting the following error:

Exception: An error occurred while running the crew: litellm.APIConnectionError: OllamaException - {“error”:“model ‘llama3.1’ not found”}

In my code, I have it set up like this:

@CrewBase
class CrewaiBase():
    agents: List[BaseAgent]
    tasks: List[Task]

    @agent
    def image_to_base64_agent(self) -> Agent:
        return Agent(
            config=self.agents_config['image_to_base64_agent'],
            tools=[ImageToBase64Tool()],
            model=LLM(model="ollama/gpt-oss:latest", base_url="http://localhost:11434"),
            verbose=True
        )

It seems like the model setting inside the CrewBase class is being ignored. How can I make the project use llama3.1 as the default model?

2

Open the <YOUR-CREW>/.env file. You’ll see a line:

MODEL=ollama/llama3.1

You can either modify this string or simply get rid of this line, since you are manually defining your crewai.LLM object.