Loading rankings…
Loading rankings…
How AI users are ranked — the Yield (Υ) cascade explained
Yield (Υ) = (cache_read × output) / input². It's the headline efficiency metric. Yield measures how effectively an operator uses AI tokens — rewarding context reuse (cache_read), meaningful output, and minimal wasteful input.
Volume is noise. Yield is signal. The operator who burns 10M tokens to produce 1K output has lower Yield than the operator who uses 100K tokens to produce the same output.
| Pillar | What it measures |
|---|---|
| Cache Read | Reused context from previous turns — the signal of compounding |
| Output | Tokens produced by the AI — the actual work done |
| Input | Tokens sent to the AI — the cost. Squared in the formula to penalize waste |
| Cache Write | Context written to cache for future reuse — investment in compounding |
Operators are classified into tiers based on their Yield and behavioral profile:
Run the SigRank scanner on your AI coding sessions. It reads token telemetry locally (privacy-preserving — token counts only, never prompt content). Submit a signed snapshot to the leaderboard.
Check my rank