Interconnects has published two new free data products to deepen coverage of the open model ecosystem: the Artifacts Hub and an Adoption Dashboard that updates daily. The tools are intended to increase transparency about how open models are adopted and used.
What the new tools do
-
Artifacts Hub: a curated view of models trending on Hugging Face that combines multiple data sources. It surfaces inference tokens via Open Router, intelligence metrics from Artificial Analysis, and Interconnects’ own adoption metrics built on top of Hugging Face data.
-
Adoption Dashboard: a living dashboard that reports download and derivative-model counts by geography and organization. The dashboard highlights the gap between the US and China and identifies growing players in the open-model ecosystem.
Data coverage and measurements
The Artifacts Hub currently covers 792 models released in the past two years. Interconnects tracks data for every model on Hugging Face, analyzes the core few thousand LLMs (their list is public on GitHub and regularly updated), and hand-selects a core few hundred models for more detailed explanation.
For popular models the Hub shows how far a model trails the frontier according to Artificial Analysis’s Intelligence Index, compares Hugging Face and Open Router adoption against similar models, displays relative adoption metric (RAM) scores that normalize downloads by time and size, and reports the VAIL similarity index for related generations. The announcement cites GLM-5.2 as an example of what the Hub can present.
Contributors and background
The Hub was developed in collaboration with Project VAIL, an AI verification startup that supported the curation work. The new products draw on data from Hugging Face, Open Router, and Artificial Analysis, and build on Interconnects’ prior work such as The ATOM Project and the Artifacts Log.
Why this matters
Interconnects says enabling the open ecosystem is central to its mission. As organizations and developers figure out how to use open models productively and cost-competitively relative to frontier models, more transparent, up-to-date data can help identify what approaches are working and where the ecosystem’s strengths lie — including observable patterns like the recurring US–China adoption differences.
Contact and acknowledgements
Interconnects invites feedback and collaboration at mail@interconnects.ai. They thank Hugging Face, Open Router, and Artificial Analysis for providing data, and Project VAIL for motivation and support in building these projects.
