genai-prices 0.0.1

Database of prices for calling LLM models.

genai-prices

Python package for github.com/pydantic/genai-prices.

This method calculates the price of API calls to a range of LLM inference providers.

Installation

uv add genai-prices

(or pip install genai-prices if you're old school)

Usage

Since this library may need to make a network call to download prices, both sync and async veriants of calc_price are provided.

Sync API Example

from genai_prices import Usage, calc_price_sync

price_data = calc_price_sync(
    Usage(input_tokens=1000, output_tokens=100),
    model_ref='gpt-4o',
    provider_id='openai',
)
print(price_data.price)

Async API Example

import asyncio

from genai_prices import Usage, calc_price_async

async def main():
    price_data = await calc_price_async(
        Usage(input_tokens=1000, output_tokens=100),
        model_ref='gpt-4o',
        provider_id='openai',
    )
    print(price_data.price)

if __name__ == '__main__':
    asyncio.run(main())

Auto Update

Both calc_price_sync and calc_price_async can be configured to auto-update by passing auto_update=True as an argument. This will cause the library to periodically check for updates to the price data.

Please note:

• this functionality is explicitly opt-in
• we download data directly from GitHub (https://raw.githubusercontent.com/pydantic/genai-prices/refs/heads/main/prices/data.json) so we don't and can't monitor requests or gather telemetry

You may also pass a custom source to auto_update to customize auto-update behavior.

At the time of writing, the data.json file downloaded by auto-update is around 26KB when compressed, so is generally very quick to download.

None-the-less, the library tries hard to avoid making a network call when the user calls calc_price_sync or calc_price_async:

• data is cached for one hour by default
• when the cached data is 30minutes old, the library will attempt to update the cache in the background
• You may pre-fetch data at program startup using genai_prices.prefetch_async() and genai_prices.prefetch_sync(), these are both sync methods which return immediately and update the cache in the background, the only difference is that calc_price_async will wait for the prefetch_async task to complete when it is first called, and calc_price_sync will wait for the prefetch_sync concurrent future to complete when it is first called.

Further Documentation

We do not yet build API documentation for this package, but the source code is relatively simple and well documented.

If you need further information on the API, we encourage you to read the source code.