mocop
Browse · Docs · Guide · Analyze

The Aggregate

Unified rankings for 10514 ranked models (2287 in the fused corpus) across 35952 public benchmarks, of which 29208 are listed in the current ranking. Refreshed daily. A predictive-evidence model combines measured results using benchmark predictive error and calibration uncertainty. Targets have no minimum benchmark or source count; only connected evidence contributes to a rating. The displayed ± is conditional capability uncertainty. Rows below are fused across reasoning-effort settings, one per model, matching the default board on /models. Human reference baselines combine benchmark-specific…

Website: theaggregate.ai

Capabilities

  • The Aggregate: LLM benchmark aggregate (MCP server): Read-only tools over The Aggregate: Aggregate Ratings fitted across public LLM benchmark leaderboards, updated daily (about_the_aggregate reports the current…
  • The Aggregate (theaggregate.ai) (Website): Unified rankings for 10514 ranked models (2287 in the fused corpus) across 35952 public benchmarks, of which 29208 are listed in the current ranking. Refreshed…
About · Contact · Privacy