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

# Performance

> Agno is optimized for agent workloads at scale.

Agent workloads spawn hundreds of instances and long-running tasks. Stateless, horizontal scalability isn't optional. It's necessary. Agno optimizes performance across three dimensions:

1. **Agent performance:** Instantiation time, memory footprint, tool calls, history management.
2. **System performance:** Async-first API, minimal memory, parallel execution, background threads.
3. **Reliability and accuracy:** Monitored through evals (covered separately).

## Benchmarks

We measured instantiation time and memory footprint for an Agent with 1 tool, run 1000 times.

<sub>Apple M4 MacBook Pro, Oct 2025</sub>

| Metric        | Agno        | LangGraph      | PydanticAI  | CrewAI       |
| ------------- | ----------- | -------------- | ----------- | ------------ |
| Instantiation | **3μs**     | 1,587μs (529×) | 170μs (57×) | 210μs (70×)  |
| Memory        | **6.6 KiB** | 161 KiB (24×)  | 29 KiB (4×) | 66 KiB (10×) |

<Note>
  Runtime performance is bottlenecked by inference and hard to benchmark accurately. We focus on minimizing overhead, reducing memory, and parallelizing tool calls.
</Note>

## Run it yourself

The benchmark code is available in this [cookbook](https://github.com/agno-agi/agno/tree/main/cookbook/evals/performance). Run it on your own machine. Don't take these numbers at face value.

```shell theme={null}
# Setup virtual environment
./scripts/perf_setup.sh
source .venvs/perfenv/bin/activate

# Run benchmarks
python cookbook/evals/performance/instantiate_agent_with_tool.py           # Agno
python cookbook/evals/performance/comparison/langgraph_instantiation.py    # LangGraph
python cookbook/evals/performance/comparison/crewai_instantiation.py       # CrewAI
python cookbook/evals/performance/comparison/pydantic_ai_instantiation.py  # PydanticAI
```

<Frame>
  <video autoPlay muted loop playsInline style={{ borderRadius: "0.5rem", width: "100%", height: "auto" }}>
    <source src="https://mintcdn.com/agno-v2/Xc0-_OHxxYe_vtGw/videos/performance_benchmark.mp4?fit=max&auto=format&n=Xc0-_OHxxYe_vtGw&q=85&s=89288701c4cb61d9d2a551fd0b5630a6" type="video/mp4" data-path="videos/performance_benchmark.mp4" />
  </video>
</Frame>

### How we measure memory

We use Python's `tracemalloc` library. First, we calculate baseline memory with an empty function, then run the Agent 1000× and measure the difference. This isolates the Agent's memory usage.

If you spot an error in our methodology, [let us know](https://github.com/agno-agi/agno/issues).

<Note>
  While we share benchmarks against other frameworks, accuracy and reliability matter more than speed.
</Note>