How Much Caffeine Is in Tea? Measured Numbers by Type

Look up the caffeine content of tea and you will find a chart within seconds. Black 47 mg. Oolong 38. Green 28. White 25. The numbers look authoritative, and they travel from one blog to the next without anyone checking where they came from. They are also close to meaningless, because a single figure per tea type cannot be right. The same leaves can give you 15 mg or 90 mg depending on three things the chart never mentions: how much leaf goes in the cup, how long it stays there, and how broken the leaf is.

That is not a rounding error. It is the difference between a cup you can drink at ten at night and one that keeps you awake.

What a controlled test actually found across all six types

A 2025 study published in Foods brewed six tea types side by side under identical conditions and measured caffeine and L-theanine by reversed-phase HPLC. Green, white, yellow, oolong, black and Pu-erh were all included, brewed at 80 °C and 100 °C for 2, 5 and 10 minutes. Because every variable except the tea itself was held constant, the comparison between types is meaningful in a way that no gathered-from-the-internet chart is.

Tea type	Caffeine (mg / 200 mL)	L-theanine (mg / 200 mL)
Black tea	14 ± 1.0	12.3 ± 2.8
Yellow tea	13.8 ± 0.2	—
Pu-erh tea	13 ± 2.9	—
Green tea	11 ± 2.1	12.5 ± 2.5
White tea	5.8 ± 0.7	—
Oolong tea	4 ± 1.6	—

Black and yellow tea came out significantly higher than green, white and oolong (p < 0.001). Pu-erh was statistically indistinguishable from black tea (p > 0.05). Oolong sat at the bottom, roughly a quarter of the black tea figure. L-theanine ran highest in green and black tea, both around 12 mg per 200 mL.

Those absolute numbers will look low if you are used to seeing 40 or 50 mg quoted for a cup of black tea, and the gap is worth understanding rather than arguing about. The study used a fixed brewing matrix, not a heaped teaspoon left in a mug for five minutes. What the numbers are good for is the ranking and the ratios: under one set of conditions, oolong delivered about a quarter of what black tea did. That ratio holds even if you scale everything up.

It also contradicts the folk rule that people repeat most often. White tea is widely sold as the low-caffeine option and oolong as a middle-of-the-road one. Under measurement, oolong was the lowest of the six and white tea was second lowest — and the difference between them was small enough that swapping one for the other changes very little.

Time does the work. Temperature mostly does not.

When people want less caffeine, the first thing they reach for is a lower water temperature. The evidence does not support that as a general strategy.

In the same study, raising the brewing temperature from 80 °C to 100 °C shifted caffeine significantly in exactly one of the six teas: oolong, which went from 4.4 to 13.5 mg per 200 mL (p < 0.05). For the other five, the temperature change did not produce a significant difference. L-theanine was stable across the temperature range in all of them (p > 0.05).

Brewing duration was a different story. Longer steeping raised both compounds across the board. Yellow tea caffeine climbed to 53 ± 16.9 mg per 200 mL at the 10-minute mark (p < 0.05) — close to four times the figure at shorter steeping.

So the practical ranking of levers is: how long, then how much leaf, then how broken the leaf is, and only in some teas does water temperature matter much. If you are adjusting one thing to cut caffeine, cut the steeping time first. If you care about water temperature for taste, that is a separate and entirely valid reason to keep adjusting it — see the tea water temperature guide for that side of it.

Broken leaf gives up more caffeine than whole leaf

This is the finding that most often surprises people, because the intuition runs the other way. Loose leaf feels like the stronger, more serious product. Tea bags feel like the weaker one.

Measured, for black tea: a cup-sized bag yielded 39 ± 9.9 mg per 200 mL, against 24 ± 7.2 mg for loose leaf of the same category (p < 0.001). The bag delivered about 60 percent more caffeine.

The mechanism is surface area. Bags are usually filled with smaller, more broken grades, and broken leaf has far more exposed surface for the same weight, so soluble compounds move into the water faster. Fannings and dust are the extreme version. A whole-leaf oolong that unfurls slowly is at the opposite end of the range, which is part of why oolong scored so low in the same test.

Curiously, serving form made no significant difference for green or white tea in this study (p > 0.05). The effect is real but it is not universal, and anyone who tells you it applies to every tea is overstating it.

The second steep carries more caffeine than the first

Gongfu brewing runs through many short infusions, and most people assume the caffeine front-loads into the first one and tapers off. A 2026 study in the Journal of Food Science (Zhang, Yusuf, Xu & Isah; doi:10.1111/1750-3841.71082) suggests that is not what happens.

Researchers brewed a green tea three times — 4 g of leaf in 100 mL at 85 °C for 2 minutes each, with a 5-minute rest between — and profiled every infusion by UPLC-MS. Caffeine was highest in the second infusion, not the first. The third was the lowest of the three. L-theanine, gallic acid and EGC all went the other way: highest in the first infusion and declining from there. Total polyphenols fell to 73 percent of the first infusion in the second, and 42 percent in the third.

In other words, the caffeine curve and the flavour curve are not the same curve. The first steep carries most of what makes the tea taste sweet and fresh; the second is where the caffeine peaks.

So does rinsing the leaves remove caffeine?

No. Not in any amount that matters. This is worth stating plainly because rinsing is standard practice in Chinese tea preparation, and it is frequently described online as a way to wash out caffeine.

Be precise about what is known. The study above did not include a rinse-and-discard treatment, so there is no direct measurement of a discarded first steep. What the data does show is that a 2-minute first infusion left enough caffeine behind for the second infusion to contain the peak amount. A rinse is a much shorter contact than that, often five to ten seconds, frequently in water that has cooled below brewing temperature. Whatever it removes is a small fraction of what a full first infusion would have taken.

Rinsing has legitimate reasons behind it. It wakes compressed tea up, it rinses dust off aged tea, and it warms the vessel. Lowering caffeine is not one of them. If you want less caffeine, shorten the steeps you actually drink rather than pouring one away.

L-theanine does not cancel the caffeine out

Tea contains L-theanine, an amino acid almost unique to the tea plant, and it gets credited with producing a calm alertness that coffee cannot. Some of that is fair. The research is thinner than the marketing.

What is established: L-theanine is present in meaningful amounts — the Foods study measured around 12 mg per 200 mL in green and black tea — and it is stable across brewing temperature, so you do not lose it by brewing cooler. A 2026 infusion study found it declined steadily from the first steep to the third.

What is not established: that it meaningfully blocks caffeine's effects in a normal cup. The ratios matter here. Tea at these measurements carries caffeine and theanine in roughly comparable amounts, and the calm-focus claim rests on interactions observed at doses well above what a cup delivers. Treat it as a plausible contributor to how tea feels rather than a proven mechanism, and be sceptical of any product copy that presents it as settled. For a fuller look at what the plant compounds in tea do and do not do, the antioxidants explainer covers the same territory with the same scepticism.

How much is too much

The European Food Safety Authority published its opinion on caffeine safety in 2015 and it remains the reference point. Its conclusions:

• Up to 400 mg per day from all sources does not raise safety concerns for adults in the general population.
• Single doses up to 200 mg, about 3 mg per kg of body weight for a 70 kg adult, do not give rise to safety concerns.
• For pregnant and breastfeeding women, habitual intake up to 200 mg per day is considered safe.

The 400 mg figure is a daily total from everything — coffee, tea, chocolate, soft drinks, and any supplement. It is not a tea allowance.

For scale, an 8-ounce cup of brewed coffee is commonly cited at around 90 to 95 mg, though that figure comes from general reference sources rather than the kind of controlled testing described above, so treat it as an order of magnitude rather than a measurement. On the tea side, the honest answer is that a cup might be 15 mg or it might be 70, and the variables above are what decide. Someone drinking strong black tea from bags, steeped long, three times a day, can reach coffee-level intake without ever touching coffee.

If caffeine keeps you awake, these are the levers

Ranked by how much they actually change the dose:

Shorten the steep. The single biggest lever, and effective in every tea tested. Going from a long steep to a short one moved yellow tea by a factor of nearly four.

Use less leaf. Linear and obvious, but people routinely use two or three times the leaf a study would.

Choose whole leaf over broken. For black tea this was worth about 60 percent. If you drink bagged black tea in the evening and sleep badly, switching to whole leaf is worth trying before giving up the habit.

Drink the first steep and stop earlier in the session. Counter-intuitive given the second-steep peak, but the total is what matters: three short infusions deliver more cumulative caffeine than one, regardless of which individual steep holds the peak.

Consider the type. Oolong measured lowest and black highest. The spread between the top and bottom of the six was roughly three and a half fold, which is real but smaller than what brewing method does. Type is the weakest of these five levers, not the first one to reach for.

Claims worth ignoring

A few things circulate widely and do not survive contact with measurement:

"White tea is the lowest in caffeine." Oolong measured lower. The difference between white and oolong was small.

"Green tea is low, black tea is high, and that is the whole story." The ranges overlap heavily depending on how you brew. A long-steeped green tea can exceed a short-steeped black one.

"Rinsing removes the caffeine." Covered above. It does not.

"Decaffeinated means no caffeine." Decaffeination reduces rather than eliminates. If you are avoiding caffeine for a medical reason, this distinction matters.

"Tea has less caffeine than coffee." Usually true per cup, and far from guaranteed. It depends entirely on the tea, the leaf form and the steeping time.

Where to go from here

If you are working through how preparation changes the cup more broadly, start with the gongfu brewing guide, which is where multi-infusion technique gets the detailed treatment this article only touches on. For how each type behaves on its own terms, the type overviews go deeper: green tea, black tea, oolong tea, white tea, yellow tea and Pu-erh. Storage does not change caffeine much, but it does change almost everything else about how a tea tastes, and shelf life is where that is covered.

The short version: no tea type is reliably low in caffeine, no chart with one number per type is trustworthy, and the dose is mostly in your hands rather than in the leaf.