Tool-Using Agent-as-Judge Evaluation

Score a CalculatorTools agent’s math answer on step-by-step clarity with numeric AgentAsJudgeEval.

Demonstrates judging responses generated by an agent using tools.

agent_as_judge_with_tools.py

"""
Tool-Using Agent-as-Judge Evaluation
====================================

Demonstrates judging responses generated by an agent using tools.
"""

from typing import Optional

from agno.agent import Agent
from agno.eval.agent_as_judge import AgentAsJudgeEval, AgentAsJudgeResult
from agno.models.openai import OpenAIChat
from agno.tools.calculator import CalculatorTools

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------
agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    tools=[CalculatorTools()],
    instructions="Use the calculator tools to solve math problems. Explain your reasoning and show calculation steps clearly.",
)

# ---------------------------------------------------------------------------
# Create Evaluation
# ---------------------------------------------------------------------------
evaluation = AgentAsJudgeEval(
    name="Calculator Tool Usage Quality",
    model=OpenAIChat(id="gpt-5.2"),
    criteria="Response should clearly explain the calculation process, show intermediate steps, and present the final answer in a user-friendly way",
    scoring_strategy="numeric",
    threshold=7,
)

# ---------------------------------------------------------------------------
# Run Evaluation
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    response = agent.run("What is 15 * 23 + 47?")
    result: Optional[AgentAsJudgeResult] = evaluation.run(
        input="What is 15 * 23 + 47?",
        output=str(response.content),
        print_results=True,
    )
    assert result is not None, "Evaluation should return a result"

Run the Example

1. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   
2. Install dependencies

   uv pip install -U agno openai
   
3. Export your OpenAI API key

   Mac/Linux

   export OPENAI_API_KEY="your_openai_api_key_here"
   
4. Run the example

   Save the code above as agent_as_judge_with_tools.py, then run:

   python agent_as_judge_with_tools.py
   

Full source: cookbook/09_evals/agent_as_judge/agent_as_judge_with_tools.py