openllm-leaderboardconditionalweight 1adapter 14 of 17

Leaderboard Rank.

The Open LLM Leaderboard adapter converts leaderboard standing into a percentile component. It counts only for models on the leaderboard, turning an external, independently maintained ranking into a directly comparable 0–100 component.

Counts only when its signals are present on the subject.

What it measures.

Percentile

Leaderboard position expressed as a 0–100 percentile.

Independent evaluation

Ranking produced by the leaderboard’s own benchmark suite.

Component keys.

openllm-leaderboard.*

Component family covering leaderboard standing.

When it counts.

Conditional — applies only to models listed on the Open LLM Leaderboard.

When signals are missing.

Unlisted models compute without this component.

How it contributes.

{
  "component": "openllm-leaderboard.*",
  "scale": "0-100",
  "mode": "conditional",
  "weight": 1,
  "contribution": "component x 1 / total weight"
}

The adapter’s components are averaged into the composite as a weighted mean: every contributing adapter’s component is multiplied by its weight, summed, and divided by the total weight of contributing adapters. The result is published as trustScores.total with a config version and timestamp so anyone can recompute it.

Questions about Leaderboard Rank.

What does the Open LLM Leaderboard adapter contribute?

It converts a model’s Open LLM Leaderboard position into a percentile component (openllm-leaderboard.*) at weight 1.

Does it apply to agents or only models?

Only models on the leaderboard. It is a conditional adapter for model subjects and does not count for other subject types.