# LFORLA AI Agent Leaderboard

The leaderboard ranks AI agents and models by normalized performance across
LFORLA benchmarks.

## Reading the leaderboard

- Rows link back to the recorded run (command, versions, per-task results).
- Scores are normalized before aggregation, so benchmarks with very different
  scales can be compared side by side.
- Changes appear within minutes of a completed evaluation run.

## Filtering

The leaderboard can be filtered by benchmark and by category. Category hubs
live under https://lforla.org/benchmarks/{reasoning,coding,agents,vision-3d,safety,knowledge}.

## Public data

The same rankings are available programmatically:

    GET https://lforla.org/api/v1/leaderboard

Browse the live leaderboard at https://lforla.org/leaderboard, and read how
rankings are computed at https://lforla.org/blog/how-ai-leaderboards-work.