Sourced snapshots · archived benchmark records
Model Leaderboard
Compare recorded agent rankings, benchmark scores, latency and prices from sourced snapshots. Every figure is shown as recorded, and a missing figure is shown as missing.
Benchmark Transparency & Methodology Standards(Click to view provenance)
Benchmark Transparency: The Arena agent evaluations represent sourced snapshots captured October 4, 2026. Net Improvement measures agent task completion delta; it is evaluated separately from coding and reasoning benchmarks.
Archived records: Benchmark scores, prices and latency on this page come from archived source records, and the conditions they were measured under are not recorded. The page opens on a curated lineup of recorded models; it is an editorial selection, not a live feed. For availability and current prices, consult the Model Directory or simulate workloads in the Cost Calculator.
Ecosystem Hubs
Explore Connected Developer Tools
Live Model Directory
286+ verified models with real-time provider list pricing and 0% token markup pass-through.
FinOps Cost Calculator
Estimate monthly token savings and compare against reseller markups.
Two-Line Migration
Repoint base URL and swap key from OpenRouter or OpenAI with zero downtime.
Press Room & Milestones
Official news releases, benchmark studies, media assets, and analyst coverage.
