What does price per quality point mean?

It is the blended cost per million tokens divided by the chosen benchmark score. A lower number means you pay less for each point of measured capability, so it normalizes raw price against how good the model actually is.

Why use a blended price?

Models price input and output tokens differently. A blended price (a weighted mix, often 3:1 input-to-output) gives a single comparable figure. You can adjust the prices to reflect your real input-output ratio.

Which benchmark should I pick?

Use MMLU for general knowledge tasks, HumanEval for code generation, and MATH for multi-step reasoning. The composite average is a reasonable default when your workload mixes all three.

Are benchmark scores comparable across models?

Roughly. Benchmarks are imperfect and providers sometimes report under different conditions, but for value ranking the relative ordering is informative. Always validate the top candidates on your own evals before committing.

Does cheapest-per-point mean I should always pick it?

Not necessarily. The cheapest-per-point model may still miss a quality floor your product needs. Use this tool to shortlist value leaders, then confirm they clear your minimum acceptable score on your tasks.

Price-per-Quality-Point Calculator

Raw token prices are misleading: a model that costs half as much but scores far lower is not a bargain. This calculator normalizes price by benchmark performance, ranking models by the cost of each quality point so you can compare on true value.

How it works

For each model the tool computes a blended price — input and output rates combined with a configurable weighting (default 3:1 input-to-output, typical of chat workloads). It then divides that blended price by your chosen benchmark score:

price per quality point = blended_price_per_million / benchmark_score

A lower result means you pay less per point of capability. Pick MMLU (knowledge), HumanEval (coding), MATH (reasoning), or a composite average of all three. Models are re-ranked instantly, cheapest-per-point first.

Worked example

The built-in models on the composite metric, blended at the default 3:1 input-to-output weight (blended = 0.75 × input + 0.25 × output):

Model	Blended $/1M	Composite	$ / point
Budget (small)	$0.26	63.3	0.0041
Open-weights	$0.33	76.0	0.0043
Mid-tier	$2.00	80.7	0.0248
Frontier	$8.75	90.0	0.0972

The budget model wins on value (lowest cost per point) — but only if a composite of 63 clears your quality floor. If your task needs 80+, the mid-tier model is the real value leader because the cheaper models simply cannot do the job.

Tips

Set a minimum acceptable score mentally before reading the ranking; value-per-point only matters above your quality floor.
Adjust the input-output weighting to match your traffic — output-heavy workloads make output-expensive models look worse.
Use this to shortlist, then confirm with your own evals and the LLM API Cost Calculator for absolute monthly spend.