Trending snapshots
Hugging Face uses its own trendingScore ordering. The site captures all 50 returned public model repositories, across all returned tasks. The score is provider-defined; it is not a measured growth percentage or benchmark.
GitHub Trending preserves every entry and the official page order from the daily and weekly captures. Period-star figures are source-labelled values, not values inferred by subtracting two snapshots. All topics and languages are included.
Popular
Hugging Face’s download-order view captures the first 50 returned repositories, including embeddings, retrieval and classic models. Its trailing-30-day query-file request counts are not unique users or adoption of commercial cloud models.
GitHub popularity is a deduplicated union of the top 100 non-fork, non-archived repositories for each of six selected topics: artificial-intelligence, machine-learning, llm, generative-ai, ai-agents and model-context-protocol. It is not all GitHub. Overlapping topic totals are not added together.
New and noteworthy
The news timeline uses official announcements in the 30 days ending 27 October 2026. Announcement, safety update, rollout and hosted availability are different dates. A newly created GitHub repository is not necessarily a newly launched project.
Editor’s selections
Reviewed profiles are selected for a documented role, an identifiable workflow and official evidence. They are alphabetical unless explicitly labelled otherwise. They are not a claim that every listed agent is top-ranked, widely adopted or suitable for every task.
What is not measured
No independent AI-product benchmarks, market-share estimates, fabricated historical charts or composite quality scores are used. Community cards and repository descriptions are unreviewed source material, not endorsements or installation recommendations.
Dated captures
These are deliberately captured snapshots. Data does not refresh continuously. Each source view gives its exact capture time and an official source link. No account connection, private user data or paid API is used to populate these public source lists.