> ## Documentation Index
> Fetch the complete documentation index at: https://agno-v2.mintlify.site/llms.txt
> Use this file to discover all available pages before exploring further.

# Gemini Basic Reasoning

> Compare Gemini 2.5 Flash with thinking disabled and a 1,024-token budget, then inspect reasoning content.

Compare Gemini 2.5 Flash with thinking disabled against a fixed 1,024-token reasoning model, then inspect the captured reasoning content.

> [!WARNING]
>
> The source labels its first agent "No Reasoning," but `gemini-2.5-flash` uses dynamic thinking when `thinking_budget` is unset. Set `thinking_budget=0` on that model to disable thinking for the comparison. See [Gemini thinking](https://ai.google.dev/gemini-api/docs/generate-content/thinking).

```python basic_reasoning.py theme={null}
"""
Basic Reasoning
===============

Demonstrates this reasoning cookbook example.
"""

from agno.agent import Agent
from agno.models.google import Gemini
from rich.console import Console


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

    # Classic reasoning test
    task = "9.11 and 9.9 -- which is bigger? Explain your reasoning."

    # Create a regular agent (no reasoning)
    regular_agent = Agent(
        model=Gemini(id="gemini-3.7-flash"),
        markdown=True,
    )

    # Create an agent with thinking budget
    reasoning_agent = Agent(
        model=Gemini(id="gemini-3.7-flash"),
        reasoning_model=Gemini(id="gemini-3.7-flash", thinking_budget=1024),
        markdown=True,
    )

    console.rule("[bold blue]Regular Gemini Agent (No Reasoning)[/bold blue]")
    console.print("This agent will answer directly without extended thinking.\n")
    regular_agent.print_response(task, stream=True)

    console.rule("[bold green]Gemini with Thinking Budget[/bold green]")
    console.print("This agent uses thinking budget to analyze the problem.\n")
    reasoning_agent.print_response(task, stream=True, show_full_reasoning=True)

    console.rule("[bold cyan]Accessing Reasoning Content[/bold cyan]")
    response = reasoning_agent.run(task, stream=False)
    if response.reasoning_content:
        console.print(
            f"[dim]Reasoning tokens used: ~{len(response.reasoning_content.split())}[/dim]"
        )
        console.print(
            f"\n[bold]Reasoning process:[/bold]\n{response.reasoning_content[:400]}..."
        )
    else:
        console.print("[yellow]No reasoning content available[/yellow]")


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

## Run the Example

<Snippet file="create-venv-step.mdx" />

**1. Install dependencies**

```bash theme={null}
uv pip install -U agno google-genai
```

**2. Export your Google API key**

```bash Mac/Linux theme={null}
export GOOGLE_API_KEY="your_google_api_key_here"
```

```bash Windows theme={null}
$Env:GOOGLE_API_KEY="your_google_api_key_here"
```

**3. Disable thinking on the regular agent**

Replace `Gemini(id="gemini-2.5-flash")` in `regular_agent` with `Gemini(id="gemini-2.5-flash", thinking_budget=0)` in the saved file.

**4. Run the example**

Save the code above as `basic_reasoning.py`, then run:

```bash theme={null}
python basic_reasoning.py
```

Full source: [cookbook/10\_reasoning/models/gemini/basic\_reasoning.py](https://github.com/agno-agi/agno/blob/v3.0.4/cookbook/10_reasoning/models/gemini/basic_reasoning.py)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.