Token Yield vs Token Count: Why Volume Lies About AI Skill
Why token yield (Υ) — not raw token count — measures AI operator skill. Your token cascade is your skill signature. The tool is the person.
When teams first start tracking AI usage, they reach for the most obvious number: how many tokens did we burn? It feels like a productivity signal. It isn't. Two operators can consume the same number of tokens and have wildly different efficiency — one compounding signal, the other just spending. Raw token count measures volume. It tells you how much you spent, not what you bought — and not how skilled you are.
The metric that actually matters is token yield (Υ). It measures the architecture of your token flow, not the volume — and that architecture is a direct reflection of your skill as an AI operator. This post explains the difference and why volume-based tracking misleads.
What raw token count measures
Raw token count is the sum of every token that crossed the wire — input plus output plus cache-read plus cache-write. It's a billing number. It tells you what you owe the API provider. It does not tell you whether those tokens produced anything useful.
Consider two operators who each consumed 2 million tokens in a week. Operator A sent 1.8M of those as fresh input, reused almost no cache, and got 200K of output. Operator B sent 400K as fresh input, reused 1.2M from cache, and got 400K of output. By raw token count they're identical. By efficiency they're opposites.
What yield measures
Yield (Υ) captures the shape of the cascade — how well your tokens compound — in a single number:
Υ = (cache_read × output) / input²The formula rewards three behaviors that define an efficient operator:
- Reusing cached context — cache_read is multiplied, so operators who build on prior turns instead of re-explaining score higher.
- Producing output — output is multiplied, so operators who extract real work from the model score higher.
- Keeping fresh input lean — input is squared in the denominator, so operators who flood the model with fresh context are penalized.
Raw token count rewards none of these. It only rewards spending. That's the core difference: volume measures how much you spent; yield measures what you got for it.
Why input is squared
The input² term is the part that surprises people, and it's the part that does the most work. Fresh input is the expensive resource in the cascade — it costs compute, it adds noise, and it doesn't compound. Squaring it in the denominator means the penalty for burning input grows super-linearly:
Double your input while keeping output and cache_read constant, and your yield drops by a factor of four — not two. The formula doesn't just discourage high input; it makes high input the single fastest way to tank your score.
This is deliberate. Operators who flood the model with fresh context — pasting entire files, repeating instructions, starting every session from scratch — are the ones whose cascades burn instead of compound. The input² term makes that behavior visible in the number.
The two operators, scored
Back to Operator A and Operator B — both at 2M tokens consumed. Here's what yield says about each:
Operator A — the burner
Input 1.8M, output 200K, cache-read 0. Yield = (0 × 200K) / 1.8M² = 0. No cache reuse means no compounding. Two million tokens, zero efficiency.
Operator B — the compounder
Input 400K, output 400K, cache-read 1.2M. Yield = (1.2M × 400K) / 400K² = 1,200. Same total tokens, but the cascade is compounding: cached context is doing work, output is flowing, input is lean.
Same volume. Opposite efficiency. Raw token count can't tell them apart. Yield can. That's why leverage and yield — not token count — are the metrics that define a skilled AI operator.
When high token usage is fine
High token usage isn't inherently bad. An operator who consumes 5M tokens but reuses 4M from cache, sends 500K as fresh input, and produces 500K of output has a high yield — they're compounding, not burning. The volume is high because the work is substantial, not because the cascade is wasteful.
The failure mode is high token usage with low cache reuse and low output — burning tokens without compounding signal. That's tokenmaxxing, and yield flags it immediately. The number doesn't care how much you spent; it cares whether your spending compounded.
The tool is the person
Here's the insight that the token-count leaderboards miss: every token the tool burns is a decision the person made. When you reuse cached context, that's your discipline. When you keep input lean, that's your restraint. When you extract high output from low input, that's your leverage. The cascade isn't the tool's behavior — it's yours.
The token-count leaderboards (clawdboard, CCgather, TrustMRT) rank tools spending money. They tell you who burned the most tokens or spent the most dollars. That's a leaderboard of hammers, not carpenters. Yield ranks the carpenter — because the cascade is the person's skill signature.
You don't need a quiz to know if someone is an AI power user. You need their token cascade. See who ranks →
Measure your cascade, not your volume
Stop tracking how many tokens you burn. Start tracking whether they're compounding:
npx sigrankAlready have token stats? Score your yield instantly →
FAQ
- What is the difference between token count and token yield?
- Token count measures how many tokens you consume — it's a billing number. Token yield (Υ = cache_read × output / input²) measures how well those tokens compound. High yield means your cached context is doing work for you. Token count tells you what you spent; yield tells you what you got for it.
- Why does input² appear in the yield formula?
- Input is squared in the denominator because fresh input is the expensive resource. Doubling your input while keeping output and cache_read constant quarters your yield. This penalizes operators who flood the model with fresh context instead of reusing cached results.
- Is high token usage bad?
- Not necessarily — high token usage with high cache reuse and high output is efficient. High token usage with low cache reuse and low output is tokenmaxxing: burning tokens without compounding signal. Yield distinguishes the two.
- How does tracking my token usage tell you about my skill?
- Every token the AI tool burns is a decision you made. High cache reuse means you build on prior context (discipline). Low fresh input means you don't flood the model (restraint). High output per input means you extract work efficiently (leverage). Your token cascade is your skill signature — the tool is the person.
Related: Yield Cascade Metric · Leverage Metric · Score Calculator