Agents use reasoning to work through problems step-by-step, backtracking and correcting as needed. Reasoning enables agents to solve complex tasks that would be impossible with a simple prompt-response. Let’s test it out by solving a simple task that gpt-4o fails at.

Create a file reasoning_agent.py

reasoning_agent.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat
from phi.cli.console import console

regular_agent = Agent(model=OpenAIChat(id="gpt-4o"), markdown=True)
reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"),
    reasoning=True,
    markdown=True,
    structured_outputs=True,
)

task = "How many 'r' are in the word 'supercalifragilisticexpialidocious'?"

console.rule("[bold green]Regular Agent[/bold green]")
regular_agent.print_response(task, stream=True)
console.rule("[bold yellow]Reasoning Agent[/bold yellow]")
reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

Run the Reasoning Agent:

python reasoning_agent.py

The regular agent fails to provide the correct answer, while the reasoning agent provides the correct answer and shows its full reasoning.

​
How does reasoning work?

Reasoning enables an Agent to think through a problem step-by-step before jumping into a response. The Agent works through different ideas, validating and correcting as needed. Once it reaches a final answer, it will validate its answer and provide a response.

​
How to use reasoning

To use reasoning, set the reasoning=True. You can also optionally provide a reasoning_model to use a different model for reasoning. When using reasoning with tools, do not use structured_outputs=True as gpt-4o cannot use tools with structured outputs.

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"),
    reasoning=True,
    markdown=True,
    structured_outputs=True,
)

​
Reasoning with tools

You can also use tools with a reasoning agent, but do not use structured_outputs=True as gpt-4o cannot use tools with structured outputs. Lets create a finance agent that can reason.

finance_reasoning.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat
from phi.tools.yfinance import YFinanceTools

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o"),
    tools=[YFinanceTools(stock_price=True, analyst_recommendations=True, company_info=True, company_news=True)],
    instructions=["Use tables to show data"],
    show_tool_calls=True,
    markdown=True,
    reasoning=True,
)
reasoning_agent.print_response("Write a report comparing NVDA to TSLA", stream=True, show_full_reasoning=True)

​
More reasoning examples

​
Logical puzzles

logical_puzzle.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = (
    "Three missionaries and three cannibals need to cross a river. "
    "They have a boat that can carry up to two people at a time. "
    "If, at any time, the cannibals outnumber the missionaries on either side of the river, the cannibals will eat the missionaries. "
    "How can all six people get across the river safely? Provide a step-by-step solution and show the solutions as an ascii diagram"
)

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

​
Mathematical proofs

mathematical_proof.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = "Prove that for any positive integer n, the sum of the first n odd numbers is equal to n squared. Provide a detailed proof."

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

​
Scientific research

scientific_research.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = (
    "Read the following abstract of a scientific paper and provide a critical evaluation of its methodology,"
    "results, conclusions, and any potential biases or flaws:\n\n"
    "Abstract: This study examines the effect of a new teaching method on student performance in mathematics. "
    "A sample of 30 students was selected from a single school and taught using the new method over one semester. "
    "The results showed a 15% increase in test scores compared to the previous semester. "
    "The study concludes that the new teaching method is effective in improving mathematical performance among high school students."
)

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

​
Ethical dilemma

ethical_dilemma.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = (
    "You are a train conductor faced with an emergency: the brakes have failed, and the train is heading towards "
    "five people tied on the track. You can divert the train onto another track, but there is one person tied there. "
    "Do you divert the train, sacrificing one to save five? Provide a well-reasoned answer considering utilitarian "
    "and deontological ethical frameworks. "
    "Provide your answer also as an ascii art diagram."
)

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

​
Planning an itinerary

planning_itinerary.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = "Plan an itinerary from Los Angeles to Las Vegas"

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

​
Creative writing

creative_writing.py
from phi.agent import Agent
from phi.model.openai import OpenAIChat

task = "Write a short story about life in 5000000 yearslife"

reasoning_agent = Agent(
    model=OpenAIChat(id="gpt-4o-2024-08-06"), reasoning=True, markdown=True, structured_outputs=True
)

reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

Was this page helpful?

Structured OutputTeams