9.11 or 9.9

Use native DeepSeek for a separate reasoning stage before a Groq response to the decimal comparison.

Demonstrates Groq with a reasoning model for numeric comparison.

11_or_9_9.py
"""
9.11 vs 9.9 Comparison
======================

Demonstrates Groq with a reasoning model for numeric comparison.
"""

from agno.agent import Agent
from agno.models.deepseek import DeepSeek
from agno.models.groq import Groq

# ---------------------------------------------------------------------------
# Create Agent
# ---------------------------------------------------------------------------
agent = Agent(
    model=Groq(
        id="qwen/qwen3.6-27b",
        temperature=0.6,
        max_tokens=1024,
        top_p=0.95,
    ),
    reasoning_model=DeepSeek(id="deepseek-reasoner"),
    markdown=True,
)

# ---------------------------------------------------------------------------
# Run Agent
# ---------------------------------------------------------------------------
if __name__ == "__main__":
    agent.print_response(
        "9.11 and 9.9 -- which is bigger?",
        stream=True,
        show_full_reasoning=True,
    )

DeepSeek retired the deepseek-reasoner alias after July 24, 2026. Before running, replace DeepSeek(id="deepseek-reasoner") with DeepSeek(id="deepseek-v4-flash") in the saved program. The final answer uses Groq in a separate call, so both DEEPSEEK_API_KEY and GROQ_API_KEY are required. See the DeepSeek V4 migration notice.

Run the Example

Set up your virtual environment

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

Install dependencies

uv pip install -U agno groq openai

Export your API keys

export DEEPSEEK_API_KEY="your_deepseek_api_key_here"
export GROQ_API_KEY="your_groq_api_key_here"

Run the example

Save the code above as 11_or_9_9.py, then run:

python 11_or_9_9.py

Full source: cookbook/10_reasoning/models/groq/9_11_or_9_9.py