Why Isn't My Llama3.2 3B Model Invoking Defined Tools in LangGraph?

I'm developing an agentic AI based on the LangGraph Academy video, but instead of using OpenAI GPT, I opted for llama3.2 3B since it's free. My expectation is that the LLM will trigger function calls to perform arithmetic computations, yet it directly computes operations (like 99 * 99) without calling the defined tools.

Below is my code

tools definition

PYTHONCopied
def multiply(a: int, b: int) -> int:
    """Multiply a and b.

    Args:
        a: first int
        b: second int
    """
    return a * b

def add(a: int, b: int) -> int:
    """Adds a and b.

    Args:
        a: first int
        b: second int
    """
    return a + b

def divide(a: int, b: int) -> float:
    """Divide a and b.

    Args:
        a: first int
        b: second int
    """
    return a / b

tools = [add, multiply, divide]

chatbot definition

PYTHONCopied
llm_with_tools = llm.bind_tools(tools)
def assistant(state:MessagesState):
    sys_message = [SystemMessage(content="You are a helpful assistant tasked with performing arithmetic on a set of inputs. You are encouraged to call a function for each of the given question!")]
    return {'messages' : [llm_with_tools.invoke(sys_message + state['messages'])]}

graph building and compiling

PYTHONCopied
builder = StateGraph(MessagesState)
builder.add_node('assistant', assistant)
builder.add_node('tools', ToolNode(tools))
builder.add_edge(START, 'assistant')
builder.add_conditional_edges('assistant', tools_condition)
builder.add_edge('tools', 'assistant')
state_graph = builder.compile()

PYTHONCopied
messages = [HumanMessage(content='What is 99 multiplied by 99?')]
resp = state_graph.invoke({'messages' : messages})
for m in resp['messages']:
    m.pretty_print()

With debugging enabled, the LLM outputs the following:

JSONCopied
{
  "generations": [
    [
      {
        "text": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 31 Jan 2025\n\nYou are a helpful assistant tasked with performing arithmetic on a set of inputs. You are encouraged to call a function for each of the given question!<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWhat is 99 multiplied by 99?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nTo calculate 99 multiplied by 99, I'll use the built-in multiplication function.\n\n```python
result = 99 * 99
print(result)

\n\nThe result of 99 multiplied by 99 is: 9801",
"generation_info": null,
"type": "ChatGeneration",
"message": {
"lc": 1,
"type": "constructor",
"id": [
"langchain",
"schema",
"messages",
"AIMessage"
],
"kwargs": {
"content": "<|begin_of_text|><|start_header_id|>system<|end_header_id|>\n\nCutting Knowledge Date: December 2023\nToday Date: 31 Jan 2025\n\nYou are a helpful assistant tasked with performing arithmetic on a set of inputs. You are encouraged to call a function for each of the given question!<|eot_id|><|start_header_id|>user<|end_header_id|>\n\nWhat is 99 multiplied by 99?<|eot_id|><|start_header_id|>assistant<|end_header_id|>\n\nTo calculate 99 multiplied by 99, I'll use the built-in multiplication function.\n\n```python
result = 99 * 99
print(result)

\n\nThe result of 99 multiplied by 99 is: 9801",
            "type": "ai",
            "id": "run-19f4ccaa-a9ba-46f1-905a-aab7d3455f02-0",
            "tool_calls": [],
            "invalid_tool_calls": []
          }
        }
      }
    ]
  ],
  "llm_output": null,
  "run": null,
  "type": "LLMResult"
}

The model simply evaluates the expression directly rather than generating structured tool call messages. Is this an issue with my configuration, or might it be a limitation of the llama3.2 3B model's ability to support function/tool calling? Any insight or guidance would be greatly appreciated.