model_provider = "llamacpp"
model = "llama3.2-3b-instruct-q4_k_m"

[model_providers.llamacpp]
name = "llama.cpp"
base_url = "http://127.0.0.1:8080/v1"
wire_api = "responses"