memory-benchmark

Run benchmark results for long-term memory backends

Get Started

1. Install

# from source
git clone https://github.com/memodb-io/memory-benchmark.git
cd memory-benchmark
pip install -e .

# install from pypi
pip install memory-benchmark

2. Setup

from memory_benchmark import CONFIG

CONFIG.llm_api_key = "YOUR OPENAI KEY"

3. Run the default benchmark.

import asyncio
from memory_benchmark import run_memory_benchmark

asyncio.run(run_memory_benchmark())

Expected Results

TODO

Support Datasets

• LOCOMO (default)

Support Methods

• Long-context LLM (default)
• Memobase
• AgenticMemory
• Mem0
• MemGPT
• Zep