Hundreds, Thousands, and 万

Once two-digit numbers are automatic, you add three new place-words and can reach ten thousand: 百 hyaku (100), 千 sen (1,000), and 万 man (10,000). The stacking logic is the same one you already know — a digit before a place-word multiplies it — but two things here genuinely surprise English speakers and are worth slowing down for: 万 introduces a ten-thousand base that English has no word for, so 10,000 is one man (一万), not ten thousand (十千); and a few of these combinations trigger sound changes (sanbyaku, roppyaku) that you'll meet in full on a dedicated page. This page gives you the machinery and flags both surprises.

百 — the hundreds

百 hyaku is 100. Put a digit in front to multiply it: 二百 (2·100 = 200), 五百 (5·100 = 500). By itself, 100 is simply 百 — you do not say 一百 (compare English, which also says "a hundred," not "one hundred," most of the time).

The three marked with ⚠ — 三百 sanbyaku, 六百 roppyaku, 八百 happyaku — are not read the "expected" way (san-hyaku, roku-hyaku, hachi-hyaku). The h of 百 shifts to b or p, and 六/八 also pick up a doubled consonant. These aren't random; they follow a tidy phonological rule explained on Sound Changes in Numbers. For now, just memorize the three irregular hundreds — they're extremely high-frequency.

千 — the thousands

千 sen is 1,000. Same rule: a digit in front multiplies it. 三千 (3,000), 六千 (6,000). And again, 1,000 by itself is usually just 千, without a leading 一 in everyday speech — though 一千 issen does appear inside larger numbers (一千万 = ten million).

Only two thousands shift: 三千 sanzen (the s voices to z) and 八千 hassen (a doubled s). Note carefully that 六千 is regular roku-sen — unlike 六百 roppyaku, six-thousand does not geminate. That mismatch between 六百 and 六千 catches everyone; it, too, falls out of the rule on the reading-changes page.

That last example — 千五百 sen-go-hyaku (1,500) — shows stacking in action: one thousand block, then a five-hundred block, read straight across. When a place is empty, you simply skip it and read only the places that have a digit: 1,001 is 千一 sen-ichi (thousand, then one — no hundreds or tens spoken), and 1,300 is 千三百 sen-san-byaku (thousand, then three-hundred — no tens or units). There is no spoken "zero" holding an empty column, the way English inserts nothing but Arabic digits show a 0.

万 — the ten-thousand that breaks English brains

Now the big one. 万 man is 10,000 — and this is where the Japanese system diverges from English at the root. English groups large numbers by thousands: one thousand, ten thousand, a hundred thousand, then a new word at million. Japanese instead has a dedicated single word for ten thousand and groups everything by that unit.

The immediate, non-negotiable consequence: 10,000 is 一万 ichi-man, never 十千 ("ten thousand"). You do not build ten-thousand out of "ten" and "thousand." 万 is its own place, and — unlike 百 and 千 — it requires the leading 一: bare 万 is not a number. You must say 一万.

Mercifully, 万 itself is phonologically calm: 三万 is a plain san-man, 八万 a plain hachi-man — none of the byaku/zen drama of the hundreds and thousands.

The second example, 一万二千 ichi-man-ni-sen (12,000), is the model for reading across the 万 boundary: the 万 block first (一万 = 10,000), then the ordinary sub-ten-thousand number (二千 = 2,000), giving 12,000. Everything below 万 is built exactly as you learned on Numbers 11–99 and in the hundreds and thousands above.

Reading a four-digit number across the places

Put it all together with a fully loaded number like 3,458:

三千四百五十八 — san-zen · yon-hyaku · go-jū · hachi

Read it place by place, largest first: three-thousand (三千), four-hundred (四百), five-ten (五十), eight (八). Each place uses the same digit-times-place logic, and you just chain them. There are no "and"s, no commas spoken, no irregular joints — only the handful of sound changes flagged above.

Why this matters for what comes next

The 万 base you've just met isn't a quirk that ends at 10,000 — it's the foundation of the entire large-number system. Above 万 come 億 (100 million) and 兆 (a trillion), each a fresh place-word four zeros further up, and English speakers routinely miscount because their instinct groups by threes. That mismatch, and a simple trick to defeat it, is the subject of Large Numbers. And the sound changes you keep seeing flagged — sanbyaku, roppyaku, sanzen, hassen — get their full, rule-based treatment on Sound Changes in Numbers.

Common mistakes

This is the defining English-transfer error. Because English says "ten thousand," learners assemble 十千. Japanese switches to a new unit — 一万 — at exactly the point English would reach ten thousand.

Unlike 百 and 千, which stand alone, 万 always takes a spoken 一 (or another digit). 一万, not 万.

Reading 三百 as regular san-hyaku is the classic slip. The high-frequency irregular hundreds are 三百 (sanbyaku), 六百 (roppyaku), 八百 (happyaku).

Because 六百 is roppyaku, learners over-generalize and say rossen for 6,000. But 六千 is the regular roku-sen — the thousands only shift on 三千 (sanzen) and 八千 (hassen).

Key takeaways

• 百 = 100, 千 = 1,000, 万 = 10,000. A digit before each place-word multiplies it.
• 100 and 1,000 can stand alone (百, 千); 10,000 must be 一万 — never bare 万, and never 十千.
• 万 groups the whole system by ten-thousands, the single biggest conceptual jump from English.
• Watch the irregular readings: 三百 sanbyaku, 六百 roppyaku, 八百 happyaku, 三千 sanzen, 八千 hassen — but 六千 stays regular (roku-sen).
• Read multi-place numbers largest-place first, chaining the blocks: 一万二千 = ichi-man-ni-sen (12,000).