Tool
Caffeine estimator
A range, never a number, from dose, method and serving size — with the things that widen the range stated plainly.
Estimated caffeine in your serving
83–158 mg
Your serving is the whole brew, so it is estimated to carry that share of the caffeine that reaches the drink from 15 g of coffee. The range is wide because the amount per gram varied that much across the published measurements it is taken from.
Estimated from dose, brew method and serving size — not a measurement of this specific cup. Species was not specified, so this uses the range measured for arabica. A blend containing robusta can hold a good deal more.
- What moves the number most
- The dose. In this model caffeine rises in step with the grams of coffee used, and adding water afterwards dilutes the drink without removing any. A smaller dose, or drinking a smaller share of the brew, is what lowers the figure. The species matters too: in the one study read that measured both side by side, green robusta averaged about 1.6 times the caffeine of green arabica, the two overlapped, and no single ratio holds from one coffee to the next.
- What this cannot tell you
- This is a model, not a measurement of your cup. The range is the span of published measurements for this kind of brew, nearly all of them made with arabica; your coffee, roast and extraction can fall outside it, and a short espresso sits towards the low end while a long one sits towards the high end. It is not medical guidance and cannot tell you how much caffeine is suitable for you.
An evidence-based range, not a measurement
This is a rough estimate, not a measurement — actual caffeine content varies by bean, roast batch and extraction. It covers the coffee only: tea, chocolate, energy drinks and other products add their own, a figure a product states per serving is counted separately, and where a product states none its caffeine is unknown rather than zero. It is not medical guidance and should not be used to judge suitability for pregnancy, medication use or any health condition. See the caffeine methodology.
What to do with the estimate
- You want to see the measurements each range was built from: Caffeine methodology
- You want to know what changes the caffeine in a cup, and what decaf still holds: Caffeine content and decaffeination
- You want a brew with a smaller dose, since the dose is what moves the figure: Coffee-to-water ratio calculator
- You are comparing one drink with another: Drink composition comparator
How it works
Caffeine in the drink is estimated as the dose multiplied by a range of milligrams per gram for the method, then shared across the brewed volume and sampled by your serving. Each range is the span of published measurements in which the study states how much coffee was used, rounded outward. Those measurements were made with arabica. For robusta, which no study read measured that way, the tool does not print one range. It makes two separate statements: at least as an arabica brewed the same way (the measured range), and cannot be more than about a ceiling (all the caffeine the richest robusta measured could hold, if all of it reached the cup). Where between the two a given robusta falls is not known. The arabica ranges are wide because the measurements are.
What the inputs mean
- Total brewed volume
- Everything the dose was brewed into — the whole pot, not the cup you are drinking, if you brewed a batch.
- Serving
- How much of that you are actually drinking.
- Species
- The measured ranges were all made with arabica, and leaving this unset gives the arabica range. No study read measured brewed robusta with the dose stated, so choosing robusta gives two separate statements instead of a measured range. The first is the arabica range, because the two species overlap. The second is a ceiling of 41 mg per gram: all the caffeine in the richest robusta measured in the papers read, brought to roasted coffee and counted as if every milligram reached the cup. That ceiling is a margin of caution, not an expectation, and it is not the top of an estimate. On average the one study that measured many green coffees of both found robusta at about 1.6 times arabica, and no single ratio holds from one coffee to the next. A blend with robusta in it can hold a good deal more than the arabica range, and a bag rarely says how much it contains.
Worked examples
A double espresso
- Dose 18 g
- Yield 36 ml
- Serving 36 ml
- Arabica
Result: Roughly 108-198 mg: a range, not a figure. That is 18 g multiplied by 6.0 and by 11.0 mg per gram.
The same shot from robusta
- Dose 18 g
- Yield 36 ml
- Serving 36 ml
- Robusta
Result: Two separate statements. At least as an arabica brewed the same way: 108-198 mg (measured brews; 18 g multiplied by 6.0 and by 11.0 mg per gram). Cannot be more than about 738 mg: everything in the richest robusta measured, if all of it reached the cup (18 g multiplied by 41.0). No study read measured brewed robusta with a stated dose, so where between the two a given robusta falls is not known.
A mug of filter coffee
- Dose 15 g
- Total brewed 240 ml
- Serving 240 ml
- Arabica, then robusta
Result: 83-158 mg as arabica (15 g multiplied by 5.5 and by 10.5). As robusta: at least as that arabica, 83-158 mg, and not more than about 615 mg (15 g multiplied by 41.0), which is a ceiling and not the top of an estimate.
A glass from a litre of cold brew
- Dose 60 g
- Total brewed 1,000 ml
- Serving 250 ml
- Arabica, then robusta
Result: 105-158 mg as arabica. As robusta: at least as that arabica, 105-158 mg, and not more than about 615 mg, a ceiling. The whole litre is put at 420-630 mg as arabica with a robusta ceiling of about 2,460 mg, and the glass is a quarter of it.
Common mistakes
Reading the midpoint as the answer
The width of the range is the finding. A cup could genuinely be at either end.
Assuming a darker roast means less caffeine, or more
In the one study read that roasted two arabicas across a wide range and brewed them, the differences were modest beside those from dose and species. A second, which measured the roasted beans of one robusta lot, found much less caffeine after its harshest roasts and does not explain why.
Comparing an espresso to a mug of filter by concentration
Espresso is far more concentrated and far smaller. What decides the total is how many grams of coffee went into each drink: a single shot from a small dose usually carries less than a mug of filter, and a double from 18 g can carry as much or more.
What the result does not mean
- This is an estimate from a general model, not a measurement of your cup. Nothing here has been laboratory-tested.
- It is not medical guidance, and it must not be used to assess safety for pregnancy, medication or any health condition.
What this does not account for
- Always shows a range, never a single number — actual caffeine content varies by bean, roast batch and extraction, not just the inputs below.
- The ranges rest on few studies: one for espresso, four for hot filter and immersion brews, two for cold brew and three for instant coffee. The methodology page lists each one and where its figure was read.
- The robusta ceiling is a bound, not a measurement. It assumes a roast loses no caffeine, that green coffee at 8 per cent moisture lost 20 per cent of its weight in the roaster, and that all the caffeine in the bean reaches the cup. Each of those is taken in the direction of more caffeine, so a real cup of robusta is expected to sit below the ceiling; how far below has not been measured.
- For filter and cold brew the studies measured caffeine dissolved, and the step to caffeine in the cup is an estimate of how much liquid the grounds keep. The espresso range is not narrowed for a short or a long shot.
- This is not medical guidance and should not be used to judge suitability for pregnancy, medication use or any health condition.
Where to go next
Background reading
Drinks and methods
Where this is taught
Become a Coffee Expert · Level 14
Telling evidence from lore, and holding what you know about coffee with the right confidence.