Current AI, founded as a non-profit partnership at the AI Action Summit in February 2025, has published the Open Source AI Gap Map v0.1 — an initial attempt to index the open-source AI ecosystem.
What was published
Gap Map v0.1 provides in-depth documentation for 421 products, broken down as follows:
- 266 software tools and libraries
- 85 models
- 50 datasets
- 20 hardware projects
These products were produced by 228 organizations and are organized into 14 categories across three stack layers: model components, product/UX, and infrastructure. In addition, the project tracks a long tail of 24,400 additional artifacts that remain uncategorized and unscored until further research and citation.
Data and license
The data behind the Gap Map is available in the currentai-org/os-ai-map GitHub repository under an MIT license. The published materials include 1,184 YAML files, along with the notebooks, schemas, and other scripts used to collect and process them. Making the dataset available under an open license enables reuse and further analysis within the license terms.
How to explore the data
Because the files are hosted on GitHub, they can be explored with tools such as Datasette Lite. The project also publishes a CSV of the GitHub repositories it is tracking: the project lists 16,185 GitHub repositories, which can be loaded into Datasette Lite for inspection.
Why this matters
Beyond providing an interactive map, the release of a machine-readable dataset gives researchers, developers, and policymakers a resource to better understand and analyze the structure and gaps of the open-source AI ecosystem. The 421 thoroughly profiled entries and the tens of thousands of additional artifacts documented by the project can support efforts to improve interoperability, reproducibility, and transparency in open-source AI.
Brief summary
With the Open Source AI Gap Map v0.1 and the accompanying MIT-licensed GitHub data release, Current AI aims to increase visibility into open-source AI projects and components. The release includes 421 detailed product entries, 1,184 YAML files, and a list of 16,185 tracked GitHub repositories, while another 24,400 artifacts wait for further research and categorization.
