AIBenchmarking site reportedly compares public and private rankings
A user highlights an adjusted cost-per-task metric that accounts for OpenRouter prices and cache hit ratios, while noting the site was still a work in progress.
TLDR
A user shared a follower-built benchmarking site, describing two features: a “benchmaxxing” leaderboard comparing public and private benchmark rankings, and an adjusted cost-per-task metric that accounts for OpenRouter prices and cache hit ratios. The user said the site was still a work in progress.
Combined views
8.8K
2 Sources, first seen 8h ago
likes
Benchmarking site reportedly compares public and private rankings
A user highlights an adjusted cost-per-task metric that accounts for OpenRouter prices and cache hit ratios, while noting the site was still a work in progress.