This repository contains an analysis of OpenRouter API usage data to understand global trends in large language model (LLM) adoption.
The project uses OpenRouter usage data to construct country-level indicators of model activity, including overall API token usage and the relative adoption of open-source models and models developed by Chinese organisations. These indicators provide insight into the evolving global landscape of AI model usage.
The following dataset is used in this analysis:
OpenRouter API Usage Data — Aggregated API usage logs from the OpenRouter platform capturing monthly token consumption by model and country.
The dataset includes information on:
- API token usage by country
- Model identifiers
- Monthly activity levels
- Model provider classifications
Additional processing steps classify models by ecosystem characteristics, including open-source status and Chinese model origin. Further details on data processing and methodology are documented within the notebooks and analysis pipeline.
Restrictions may apply to the data that support the findings of this analysis. Data received from the private sector through the Development Data Partnership are subject to the terms and conditions of the data license agreement and the "Official Use Only" data classification.
These data are available upon request through the Development Data Partnership: https://datapartnership.org/
Licensing and access information for all other datasets are included in the documentation.
This project is licensed under the Mozilla Public License.
See the LICENSE file for details: MPL-2.0
The World Bank Data Lab template used to create this project maintains a Code of Conduct to ensure an inclusive and respectful environment for everyone. Please adhere to it in all interactions within our community.
See: docs/CODE_OF_CONDUCT.md