Logseq/Entity/AI/Leaderboard
AI Leaderboard
- In this garden, AI Leaderboard pages model a current ranked list of Logseq/Entity/AI/Model pages along one dimension — Intelligence, Speed, Cost per Task — taken from a named public board such as ArtificialAnalysis.
What counts as an instance
Naming and links
Finding and deduplicating
- Search the source URL, then
Source/Leaderboard/<Dimension>. One page per source-and-dimension.
How a refresh works
- Read the live board. Rewrite the list in descending desirability — highest Intelligence or Speed first, lowest Cost per Task first.
- Set the heading's "as of" date to the check day.
- Every row is a garden model page plus its score. A model already in the garden is linked at its canonical path. A top-row model that is missing is created as a Logseq/Entity/AI/Model (stub is enough) with its Logseq/Entity/AI/Model/Doc/API when the official URL is known.
- Variants of the same release (max, xhigh, high) collapse to the garden's one model page. The row names the variant the score belongs to in a child line.
- Models the garden already tracks appear even when they are not in the source's top handful, as long as a score for that dimension is known.
Frontmatter
- Mark instances with
logseq-entity:: [[Logseq/Entity/AI/Leaderboard]]. - Shared frontmatter conventions live on Logseq/Frontmatter.
Page shape
- Headings, not tables or charts. The H1 is the dimension and the source link, plus
as of [[YYYY-MM-DD Day]]. - Each rank is an H2:
## [[Provider/Model/…]] - <score>. A child line holds the source's display name when it differs from the garden page. - Scores run from most desirable to least.
Relationship to other types
Examples in this garden