Interconnects has launched two free data tools for tracking the open model ecosystem. The Artifacts Hub indexes 792 models released in the last two years, pulling inference token data from OpenRouter, intelligence benchmarks from Artificial Analysis, and download metrics from Hugging Face. The Adoption Dashboard maps download counts and derivative model numbers by geography and organization, with explicit focus on the US-China gap.
The scope behind these tools is worth understanding. Interconnects tracks every model on Hugging Face, maintains a public GitHub list of several thousand curated LLMs, and hand-selects a few hundred for deeper analysis. The methodology builds on work from The ATOM Project, including a published arXiv report from April 2025. That infrastructure is now surfaced publicly, not paywalled.
The original post is short but the links carry the weight. The Adoption Dashboard's geographic breakdown of derivative models is the sharpest signal yet on where open-weight momentum is actually building. Read it to understand the data sources, then go straight to the dashboards.
[READ ORIGINAL →]