Commonwealth Towers: 45 Resales At S$2,295 psf Over Four Quarters — But The Median Just Slipped To S$2,197 psf

A project can look busy and flat at the same time. Commonwealth Towers is a reasonable example of that. The caveats keep coming in, quarter after quarter, in sizeable numbers for a single development. Yet the line joining those quarterly medians ends up close to where it started two years ago.

According to URA Data Service caveats, the 99-year project in District 3 recorded 45 resales over 3Q2025 to 2Q2026, at a median of S$2,295 psf. It has 86 years of lease left. Across the eight quarters we pulled, the quarterly median moved -0.6%. It sat at S$2,211 psf in 3Q2024 on eight sales. In 2Q2026 it was S$2,197 psf on eleven. In between, the line went up and came back down. Here is what stood out to us when we worked through the tables.

The eight-quarter line ends close to where it began, but it did not get there in a straight line

The quarterly medians peaked in the middle of the window. 3Q2025 came in at S$2,319 psf on 15 sales, the busiest quarter of the eight. 4Q2025 was marginally higher at S$2,327 psf on 10 sales. Then 1Q2026 recorded S$2,265 psf on 9 sales, and the latest quarter came in lower again. So the -0.6% move across the full stretch hides a run-up and a pullback of a few percent each way.

The percentile columns are worth a look too. In 3Q2024 the middle half of sales sat between S$2,178 psf and S$2,296 psf. In 2Q2026 the same band ran from S$2,156 psf to S$2,345 psf. That is a wider spread around a lower midpoint. Quarterly counts here run from 8 to 15. At that size, two or three large units can shift a median on their own. The table does not control for floor, stack or unit size, so some of the movement is mix rather than price.

The quarter-by-quarter numbers, and the same series drawn as a line, are below.

The smallest units did most of the trading, and the largest ones carried the highest psf

Breaking the 45 resales down by size changes the picture. Units under 50 sqm were the most traded band by a distance. There were 19 of them, at a median of S$970,000, or S$2,156 psf on a median area of 463 sqft. That band alone is close to half the project's resale activity over the four quarters.

The psf ladder runs the other way. The 50 to 69 sqm band logged 7 resales at S$2,381 psf, with a median price of S$1,640,000. The 70 to 89 sqm band had 12 resales at S$2,339 psf, median price S$1,779,000. The largest band, 90 to 119 sqm, recorded 7 resales at S$2,465 psf, with a median price of S$2,600,000 on 1,055 sqft.

So the project's headline median sits between those bands. It reflects whatever mix of sizes happened to transact. The small-unit band is well populated; the two larger bands have seven sales each. Read those two with that in mind. The size breakdown is below.

Rents stayed within a narrow band while the number of contracts swung sharply

The rental side is a larger sample than the resale side. URA rental contracts at Commonwealth Towers numbered 84 in 2Q2026, at a median of S$7.56 psf a month. The median monthly rent was S$3,750. The three quarters before that sat in the same narrow range, between S$7.56 and S$7.69 psf a month.

Contract volumes moved far more than the rates did. 3Q2025 recorded 121 contracts, the highest of the four quarters shown. 4Q2025 dropped to 57, then 1Q2026 came in at 69. Median monthly rents over those quarters ranged from S$3,600 to S$3,750. The counts here reflect approvals filed in each quarter, so lease cycles matter as much as anything else.

Set against the project's resale median, our engine measures a gross yield of about 4.0%. That is a measured figure from recorded rents and recorded resales. It is not a return on any particular unit, and it takes no account of costs. The quarterly rental table is below.

Against its district and its closest neighbours, the project sits in the upper middle of the range

Over the same four quarters, the District 3 resale median was S$2,216 psf from 614 sales. Commonwealth Towers sits +3.5% against that. That is a whole-district comparison, so it mixes tenures, ages and sizes together.

Within 1 km, the nearest comparison is Queens Peak, 137 m away. It recorded 39 resales at a median of S$2,242 psf. Stirling Residences, 526 m away, was the busiest project in the radius with 116 resales at S$2,384 psf. Margaret Ville logged 23 resales at S$2,251 psf. Those three are all 99-year leasehold.

The freehold projects nearby traded lower on psf. Alexis recorded 14 resales at S$1,883 psf, and Queens 13 at S$1,842 psf. The Anchorage had 7 at S$2,024 psf. Two entries are too thin to read much into. Queensway Tower logged 3 resales at S$1,402 psf, and Shepherd's Hill Estate just 2, at S$2,535 psf. Age and unit mix differ across all of these. The full neighbour table is below.

The top five prices all came from larger units, and mostly from the higher floors

The highest recorded resale in the window was S$2,730,000. That was in 3Q2025, for a 1,076 sqft unit on the 41 to 45 storey range. It worked out to S$2,536 psf. The second highest was S$2,712,000 in 1Q2026, for 1,033 sqft on the same storey band.

That second sale carries the highest psf in the table, at S$2,625 psf. It shows how price and psf can rank differently once areas differ. Two other sales cleared S$2.5M: S$2,700,000 in 4Q2025 on the 36 to 40 band, and S$2,588,000 in 3Q2025 on the 16 to 20 band.

Every entry in this table is 1,033 sqft or larger. The storey ranges skew high, though the 16 to 20 band appears too. Five sales is a small set, and these are the top of the distribution by definition. They are not a guide to the typical transaction. The top five are below.

Our own estimate at this project came within a median of 3.0% of the later recorded prices

We also test our estimate against sales that happened after it was produced. At Commonwealth Towers, as of 4Q2025, it was tested against 39 later sales. The median error was 3.0%. Of those sales, 26% landed inside the stated range.

Those two numbers say different things. The median error measures how far off the midpoint was, half the time better and half the time worse. The hit rate measures how often the recorded price fell inside the band we published. A tight band will post a lower hit rate at the same accuracy. Both are backward-looking, and both come from recorded transactions.

Everything here is indicative and drawn from caveats and rental filings. None of it is a formal valuation of any home in the project.

A few things the figures do not settle. The eight-quarter table cannot separate price movement from mix, because the counts per quarter are small and the size bands trade unevenly. The widening gap between the 25th and 75th percentiles in the latest quarter may be mix too. The rental counts swung hard while the rates barely moved, so the next quarter's contract volume is worth checking against the 84 recorded in 2Q2026. Whether the larger-unit bands keep posting the higher psf is the other thing to watch, given only seven sales in each.

One quarter's median rarely tells you much on its own, least of all at a single project. What it is worth is the context around it: the eight-quarter shape, the size mix underneath, the rents filed in parallel, and the district line it is measured against. Read together, those give you a direction. Read alone, a number like S$2,197 psf is just one point on a line that has already moved both ways.

Method: URA resale caveats at the project (developer sales and sub-sales excluded) by the quarter of the sale month. Size bands by caveated floor area. Rents are URA rental contracts at the project, and the gross yield is the mean of the quarterly median rents per square foot, annualised, over the four-quarter resale median per square foot, before any cost. Lease left is the tenure less the years since the lease started, as recorded. The estimate accuracy is this site's backtest: the engine run as of the quarter named, compared with the sales since. Private figures are from URA Data Service caveats, HDB figures from data.gov.sg resale records, both loaded weekly by this site's ingest. Medians are of the non-excluded transactions in the period named. Sales flagged by the hygiene rules (related-party transfers, bulk deals, outliers) are left out. Every figure is indicative and drawn from recorded transactions, not a formal valuation.