Yield measures how efficiently you use AI tokens. It's calculated as (cache_read × output) / input². Higher Yield means you're reusing cached context and producing output efficiently relative to what you send to the model.

What are the four pillars of the token cascade?

The four pillars are Cache Read (tokens served from cache — free, instant, smart), Output (tokens the model produced for you), Input (tokens you sent to the model — your cost), and Cache Write (tokens written to cache for future reuse). The cascade string shows the ratio between these pillars.

Leverage measures how much output you generate relative to your input. A high Leverage score means you're getting more results per token spent — a sign of effective prompting.

10xDEV is the base-10 logarithm of your Leverage score. It's a normalized scale that makes it easier to compare operators across different volume tiers.

How do I get my rank?

Visit signalaf.com/score to enroll and submit your token telemetry. SigRank will compute your Yield, Leverage, 10xDEV, and place you on the AI User Leaderboard with a class tier designation.

How it works

Q: Why does SigRank use Yield instead of raw token volume?

Volume is noise. Anyone can spend tokens. Yield rewards skill — a power user with 10K tokens and high cache reuse can beat someone with 10M tokens and no cache strategy. Yield measures efficiency, not spend.

The 60-second explainer. The full methodology lives on signalaf.com.

What is Υ Yield?

Υ Yield measures how efficiently you use AI tokens. It's your cache reads multiplied by your outputs, divided by your inputs squared.

Υ = (cache_read × output) / input²

In plain English: how efficiently do you reuse cached context and produce output relative to what you send to the model. Higher is better.

Why Yield, not volume?

Volume is noise. Yield is signal.

Anyone can spend tokens. The question is whether you're using them efficiently. A power user with 10K tokens and high cache reuse can beat someone with 10M tokens and no cache strategy. Υ Yield rewards skill, not spend.

The four pillars

Cache Read

Tokens served from cache — free, instant, smart.

Output

Tokens the model produced for you — your results.

Input

Tokens you sent to the model — your cost.

Cache Write

Tokens written to cache for future reuse — investment.

The cascade string (e.g., 1.5×10.2×24.5) shows the ratio between these pillars. A high cache-read ratio is the hallmark of a power user.

Ready to measure yourself?

Get your Υ Yield, your rank, and your operator class on SigRank.

Check my rank

Want the full methodology? Read the docs on signalaf.com

Loading rankings…