What 68 new homes actually cost to build

Figures

Single dwelling — net construction cost, before builder's margin and GST

Deciles on the screened population: p10 A$527,909, p90 A$2,517,373.

Where the 63 screened projects sit

By state, screened population

State counts below N=5 are single or near-single observations. They are listed for completeness, not as a state benchmark.

Split between structure and finishes

Each estimate carries a lock-up value, meaning the project brought to a weatherproof shell with roof, external walls, windows and external doors in place, before internal fitout.

The remaining 49% to 58% is internal fitout, services completion, external works and site preliminaries. The builder's margin is not inside it — the margin sits on top of the whole total, as set out in the basis note above.

Cost per square metre

Basis: gross floor area as recorded in the delivered bill of quantities for each job, divided into the net construction total, before builder's margin and before GST, screened as set out below. On the same packs the median GST-inclusive rate after margin is roughly 1.27 times the figures in this table.

Floor areas behind the single-dwelling rate

How the rate population was screened. Of the 106 completed jobs, 81 have a delivered bill of quantities carrying a floor area. Nineteen were removed, in this order:

• Four rows with a recorded area below 60 m². At that size the recorded area is partial rather than whole-building, and the implied rates prove it — two of the four imply A$19,654/m² and A$21,297/m², which no Australian build supports. The area is wrong on those rows, not the total.
• Two rows whose implied rate fell outside A$1,500 – A$10,000/m². One implies A$956/m², below any credible build rate. One implies A$14,069/m².
• Thirteen rows where the rate stored in the pack and the rate computed from the construction total disagree by more than 10%. Where those two disagree, one of the inputs is wrong and there is no way to tell which from the record alone.

That leaves 62 residential projects, 40 of them single dwellings.

The median does not depend on which rate definition is used. The delivered pack carries a stored rate per square metre, and the construction total divided by the recorded area produces a second one. Across the 75 rows where both exist and the area is whole-building, the two medians are A$4,045 and A$4,001 — 1.1% apart. Fifty of those 75 agree to within half a percent. The published figure survives either definition, which is what makes it safe to quote.

One caveat on the N=40 population. One of the 40 is the earlier run of the superseded duplicate pair named in the screening section above, which the totals table excludes. Removing it moves the median from A$4,013 to A$4,015, a shift of 0.06%. It is left in so that the rate population matches the finding it was harvested from, and named here so the difference is visible.

Two of the 40 records classified single dwelling carry a dwelling count of two in their delivered pack. Their rates are A$3,788/m² and A$4,497/m², both inside the published range, so they do not distort the median. The classification and the pack disagree on those two rows and neither has been corrected here.

Other residential types — case rates, below publishable N

Three further residential types survive the screen but with too few observations to carry a median. Each rate is listed individually.

Read these as three and two individual projects. Two or three observations have no middle.

Granny flat and secondary dwelling rates are withheld. Three rows survive the screen, but two of the three record a floor area of exactly 60.0 m², which is the screen floor itself. An area landing exactly on the threshold twice reads as a placeholder rather than a measurement. The third, at 163.1 m², implies A$2,457/m². One observation is not a figure, so no granny flat rate is published.

Basis and screening

• Source: EstiFlow production job records, filtered to jobs with status COMPLETED and a construction total above zero. That filter returns 106 jobs across all dwelling types. 68 of them are classified single dwelling.
• Completion dates: 30 April 2026 to 24 August 2026.
• Delivery evidence: 67 of the 68 single-dwelling jobs have deliverable files sealed against the job record with content hashes. 100 of the 106 jobs across all types shipped a bill of quantities workbook, 98 shipped a cost report PDF, 94 shipped subcontractor packs.
• Currency: Australian dollars, net of builder's margin and net of GST. See the basis note at the top of this article.
• Five rows removed to produce the screened population, each named:
• One NSW project recorded at A$17,158,160. Its internal split reads 26% lock-up against a 45% median, and its delivery record shows seven automated repairs including a seeded floor area and a re-applied commercial projection. The number is not corroborated by its own components. Removed and flagged.
• One NSW strata unit recorded at A$245,385. The job scope is remedial work to a single apartment in a Class 2 building, not a new home. Wrong scope class for this population. Removed.
• Two rows for the same NSW site, completed two days apart, recorded at A$1,010,933 and A$2,853,643. The second is a re-run on a later pipeline version. They differ by a factor of 2.8 and cannot both be correct. Neither is published. Both removed and flagged.
• One of a pair of rows for the same NSW site, completed seven days apart, at A$728,299 and A$814,327. The earlier run is superseded by the later one. The earlier is removed, the later retained.

What moves the number

The lower quartile is not all new-build scope. The dataset has no field separating a full new build from an alteration, an addition, or a second dwelling on an existing lot, so all of those sit in the same population. The site labels show it directly: the fourth-lowest figure, A$437,135, carries an apartment unit number in its site label, and the second-lowest, A$228,261, carries a subdivided-lot suffix. A reader benchmarking a full detached new build should treat the lower quartile as containing partial-scope work and read from the median up.

State is a weaker driver than the spread within a state suggests. The NSW median of A$903,352 and the VIC median of A$929,686 sit 2.9% apart, and the four states with more than three observations span medians from A$586,531 to A$1,077,660. NSW alone spans A$204,627 to A$5,351,909. The spread inside one state is an order of magnitude wider than the spread between state medians. Location moves a total through site access, ground conditions and council conditions, not through a state multiplier.

The upper tail is a different product under the same label. Seven of the 63 screened projects sit above A$2.5M, and the highest is 26 times the lowest. They are single dwellings by classification and nothing like a project-builder home by construction. The dataset does not record which features drive each of them, so no attribution is offered here. What the spread does establish is that the median is the number to quote for a typical detached home, and that the mean of A$1,344,468 is pulled up by that tail and is the less useful statistic.

Fitout carries more than half. The 45% median lock-up share means that at the point the shell is weathertight, less than half the money is spent. Where a build overruns, it is more often in the second half — joinery, tiling, stone, appliances, glazing, external works — than in the frame and roof.

Lot conditions the totals do not show. The dataset holds no field for slope, soil classification, bushfire attack level, flood overlay, heritage listing or demolition extent. Each of these is priced inside individual estimates and none is separable across the population. Two projects with identical floor plates in the same suburb can differ materially on those grounds alone, and this table cannot show by how much.

What's excluded

These are construction cost totals. They do not include:

• Land acquisition, stamp duty and finance costs.
• Statutory and authority fees. These are enumerated as line items in the delivered packs and are never priced into a total. For a NSW project the set enumerated is the NSW Home Building Compensation Fund contribution where applicable, the Long Service Levy where applicable, and council contributions where conditioned. For VIC it is Victorian Building Authority and building permit context, plus local council contributions where conditioned. For QLD it is QBCC Home Warranty where applicable, the QLeave levy where applicable, and local infrastructure charges where conditioned. For WA, SA and TAS the packs enumerate building permit and local government requirements where applicable. All of these are listed and excluded from the totals above, because the amounts are set by the authority and by the project's conditions of consent rather than by the builder.
• GST and the builder's margin. Every figure above is before both. This is a change from an earlier version of this article, which described the totals as merely excluding GST and was incomplete. Multiply by roughly 1.27 for a margin-inclusive, GST-inclusive figure.
• Professional fees. Architectural, engineering, certification, surveying and town planning fees are outside the construction total.
• Loose furniture, appliances not fixed, landscaping beyond what the approved drawings show, and any work not documented in the plans submitted.
• Contingency held by the owner. Where a contingency appears inside an estimate it is disclosed as a line, not folded silently into the total.

About this dataset

EstiFlow is not a registered quantity surveying practice, and the documents behind these figures are not certified quantity surveying reports. They are cost estimates, bills of quantities and estimating packs produced from the architectural drawings and supporting documents supplied for each project. Each figure above traces to a job record with sealed, content-hashed deliverable files. Where a figure could not be corroborated against its own components or against a duplicate run of the same site, it has been named and removed above rather than published.

The population is self-selected. It is the work EstiFlow was engaged to estimate, weighted toward NSW, toward detached housing, and toward projects at development-application stage. It is not a random sample of Australian residential construction.

Figures dated 2026-09-01.