March 30, 2026
By the HalfKey team
When monthly furnished beats Airbnb in Tokyo
At 21 days, the Airbnb price beats the monthly. At 30, the monthly wins by an enormous margin. The interesting band is in between, and the answer turns on a setting in someone else's pricing engine.
Booking a 21-night furnished stay in Tokyo, you will probably price it wrong. You see Airbnb at ¥12,000/night, multiply by 21, and get ¥252,000. You feel like you have a number. Then you see a "monthly mansion" listing at ¥150,000. Divide by 30, get ¥5,000/night, and the monthly wins by an enormous margin.
Both numbers are wrong. The first is missing the Airbnb monthly discount you might actually get on a 21-day stay (about 20% off, sometimes 30, sometimes nothing, depending on the host). The second is missing several line items. Departure cleaning, the utility flat fee, and the fact that you pay for 30 days whether you stay 21 or not.
Before deciding which option wins, you have to assemble the actual price of each one.
Take a mid-tier furnished studio in central Tokyo. Rent: ¥150,000 for the month. The operator charges a ¥50,000 departure cleaning, billed once at checkout regardless of stay length. Utilities come bundled as a ¥15,000 flat fee per month. They call it 光熱費 (kōnetsuhi — literally "light-and-heat fee," the catch-all line for utilities).
In a midterm-furnished setup, the operator pays the building's metered bills. You pay the operator. The flat fee is how they avoid metering you separately, which is administrative pain.
A true 30-day stay costs ¥215,000 total, or ¥7,167/night. That's the reference number.
The cleaning fee is the line worth a second look. The operator hires a contractor. The contractor charges the operator about ¥18,000 for a unit that takes two people three hours. The remaining ¥32,000 is operator margin on cleaning.
Cleaning is the only fee that scales with turnover rather than stay length. A 90-day stay needs the same departure cleaning as a 30-day stay, but produces three times the revenue against the same cleaning cost. Operators love long stays. The ¥50,000 cleaning fee is, partly, a soft penalty on the short ones.
Now run the math on a 21-day stay.
The operator's pricing is not "¥150,000 for a fraction of a month." It is: you take a month. Unit blocked off to other guests for the full month. Rent ¥150,000, plus ¥50,000 cleaning, plus ¥15,000 utility. Total: ¥215,000 across 21 days of usage, or ¥10,238 per night.
The 21-day Airbnb at ¥12,000/night with no monthly discount applied totals ¥252,000, or ¥12,000/night. (Airbnb's monthly tier kicks in at 28 nights, not 21.) The monthly wins, but only because the Airbnb host wasn't generous.
Apply a 20% monthly discount on the 21-day stay (some hosts will do that ad-hoc, preferring a long stay to a turn-down at a date threshold). Airbnb's effective rate becomes ¥9,600/night. ¥201,600 total. Airbnb wins on the same booking by ¥13,400.
Same unit, same nights, same cleaning effort. Two completely different answers, depending on a single toggle in someone's pricing settings.
At 28 nights, Airbnb's monthly discount kicks in mechanically. The question stops being about host generosity. The monthly is ¥215,000 across 28 days, or ¥7,679/night, against Airbnb's ¥9,600. Monthly wins.
At 60 days the gap is brutal. Two months' rent (¥300,000) plus two months' utility (¥30,000) plus one cleaning (¥50,000): ¥380,000 total, or ¥6,333/night, against Airbnb's ¥576,000 (60 × ¥9,600). The monthly saves ¥196,000.
Past 30 days the line opens up and never closes. Below 28 days the answer depends on whether the Airbnb host enabled their monthly discount. The interesting band is 21 to 28 days, where the answer turns on a setting in someone else's PMS.
There is a complication, and it is the one that broke the math at the start. At 21 days, you may not actually be allowed to book the monthly.
A lot of Tokyo midterm-furnished operators will not take stays under 30 days. Under 30, you are an Airbnb or hotel customer, not a monthly one. Different building, different license, different pricing engine. The building is licensed under the residential framework, not ryokan-gyō-hō. The operator's PMS recognizes monthly rent, not nightly rates.
The "21-day monthly" math is hypothetical for most buildings. The actual choice in front of most readers is 21 days at Airbnb or 30 days at a monthly. That collapses the crossover question. Do you stretch your stay to 30, or cut it down to under three weeks?
Both pricing engines respond to the same underlying reality.
The marginal cost of housing a guest for 30 nights is barely higher than the cost of housing them for 25. The revenue from 30 nights is meaningfully higher than the revenue from 25. So both sides would much rather sell you 30.
Airbnb's monthly discount is the polite version. The monthly mansion's "we won't take you for less than 30" is the impolite version. Same answer, different surface.
The unit you sleep in does not care whether your stay rounds up or down. The pricing engines on both sides do.