Introducing our Artifacts Hub and Adoption Dashboard
Interconnects launches The Artifacts Hub, a curated dashboard tracking 792 open models from Hugging Face, integrating inference token data from Open Router, intelligence rankings from Artificial Analysis, and custom adoption metrics The Adoption Dashboard provides daily-updating geographic and organizational breakdowns of model downloads and derivative models, with a focus on the US-China adoption gap The Artifacts Hub features Relative Adoption Metric (RAM) scores for time-size normalized downl
Analysis
TL;DR
- Interconnects launches The Artifacts Hub, a curated dashboard tracking 792 open models from Hugging Face, integrating inference token data from Open Router, intelligence rankings from Artificial Analysis, and custom adoption metrics
- The Adoption Dashboard provides daily-updating geographic and organizational breakdowns of model downloads and derivative models, with a focus on the US-China adoption gap
- The Artifacts Hub features Relative Adoption Metric (RAM) scores for time-size normalized downloads, VAIL similarity indices for model generations, and Intelligence Index comparisons against frontier models
- These tools build on prior Interconnects work including The ATOM Project and monthly Artifacts Log round-ups, developed in collaboration with AI verification startup Project VAIL
- The initiative aims to increase transparency in the open model ecosystem to help practitioners identify cost-competitive alternatives to frontier models
Why It Matters
Open model tracking has been fragmented, making it difficult for practitioners to assess which models are gaining real adoption versus hype. These dashboards provide the first consolidated, daily-updating view of open model adoption dynamics, intelligence gaps, and geographic distribution—critical data for anyone making sourcing or deployment decisions in the open AI ecosystem.
Technical Details
- The Artifacts Hub covers 792 models released in the past two years across text-focused LLMs and multimodal generative models, hand-selected from a broader list of a few thousand LLMs tracked on Hugging Face
- Metrics include Artificial Analysis Intelligence Index (frontier gap measurement), Open Router inference token volumes, Hugging Face download counts, RAM scores (time-size normalized adoption), and VAIL similarity index for generational lineage
- The Adoption Dashboard tracks download and derivative model numbers by geography and organization, with daily updates highlighting the US-China adoption gap
- Data sources are integrated from Hugging Face, Open Router, Artificial Analysis, and Project VAIL, with the core LLM list publicly maintained on GitHub
Industry Insight
- The US-China adoption gap dashboard fills a critical intelligence need for organizations evaluating geopolitical risk and supply chain diversification in open model sourcing
- RAM scores and time-size normalized metrics offer a more nuanced adoption signal than raw download counts, helping practitioners identify models with sustained rather than viral uptake
- The collaboration between an analytics outlet (Interconnects), a verification startup (VAIL), and infrastructure platforms (Hugging Face, Open Router, Artificial Analysis) signals a growing ecosystem of open model observability tools that practitioners should monitor for deployment decision support
Disclaimer: The above content is generated by AI and is for reference only.