Fast Reasoning

Time Llama 3.3 70B on Groq answering a math problem and check its reasoning depth.

For free and developer tiers, Groq will shut down llama-3.3-70b-versatile on August 16, 2026. Replace it with openai/gpt-oss-120b before that date. See Groq deprecations.

fast_reasoning.py

"""
Fast Reasoning
==============

Demonstrates this reasoning cookbook example.
"""

import time

from agno.agent import Agent
from agno.models.groq import Groq
from rich.console import Console


# ---------------------------------------------------------------------------
# Create Example
# ---------------------------------------------------------------------------
def run_example() -> None:
    console = Console()

    # Test task requiring reasoning
    task = "What is 23 × 47? Show your step-by-step reasoning."

    console.rule("[bold cyan]Groq Fast Reasoning Demo[/bold cyan]")

    # Test with Llama 3.3 (reasoning capable)
    console.print(
        "\n[bold blue]Llama 3.3 70B Versatile (Reasoning Capable)[/bold blue]"
    )
    try:
        start = time.time()
        agent_deepseek = Agent(
            model=Groq(id="llama-3.3-70b-versatile"),
            markdown=True,
        )
        response = agent_deepseek.run(task, stream=False)
        end = time.time()

        console.print(response.content)
        console.print(f"\n[dim]Response time: {end - start:.2f}s[/dim]")

        if response.reasoning_content:
            reasoning_len = len(response.reasoning_content.split())
            console.print(f"[dim]Reasoning depth: ~{reasoning_len} words[/dim]")
    except Exception as e:
        console.print(f"[red]Error: {e}[/red]")

    # Test with Llama for comparison
    console.print("\n[bold green]Llama 3.3 70B (Standard Mode)[/bold green]")
    try:
        start = time.time()
        agent_llama = Agent(
            model=Groq(id="llama-3.3-70b-versatile"),
            markdown=True,
        )
        response = agent_llama.run(task, stream=False)
        end = time.time()

        console.print(response.content)
        console.print(f"\n[dim]Response time: {end - start:.2f}s[/dim]")
    except Exception as e:
        console.print(f"[red]Error: {e}[/red]")


# ---------------------------------------------------------------------------
# Run Example
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    run_example()

Run the Example

1. Set up your virtual environment

   Mac

   uv venv --python 3.12
   source .venv/bin/activate
   

   uv venv --python 3.12
   .venv\Scripts\activate
   
2. Install dependencies

   uv pip install -U agno groq
   
3. Export your Groq API key

   Mac/Linux

   export GROQ_API_KEY="your_groq_api_key_here"
   

   $Env:GROQ_API_KEY="your_groq_api_key_here"
   
4. Replace Llama 3.3

   Replace llama-3.3-70b-versatile with openai/gpt-oss-120b in the saved file.

5. Run the example

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

   python fast_reasoning.py
   

Full source: cookbook/10_reasoning/models/groq/fast_reasoning.py