Mini-split heat pumps compared on certified ratings

Why can’t this data separate single-zone from multi-zone?

Because the federal file has no zone-count field. A single-zone mini-split is one outdoor unit driving one indoor head.

A multi-zone system is one outdoor unit driving several, often of different sizes, frequently running at part load against each other.

They are different machines doing different jobs, and they do not rate the same: sharing one compressor across several heads costs efficiency, and the certified numbers reflect that.

ENERGY STAR does not publish which is which: Its air-source heat pump file has one product type — "HP - Mini or Multi Split" — covering both, and no zone-count column anywhere in the record.

The industry register that does carry that field is the AHRI Directory, whose terms of use prohibit its contents being reproduced or entered into a database for derivative use, and whose licensing FAQ answers this exact question with "No".

Public browsability is not permission. So the field is unavailable to us and we do not have it.

HyreHVAC analysis — what that costs this page, stated plainly. Because multi-zone equipment generally rates below a single-zone version of the same outdoor unit, mixing them in one band pulls the band median down and makes multi-zone entries look better relative to their true peers than they are.

The effect is concentrated in the middle and upper capacity bands, because the smallest bands are necessarily single-zone: a 1,006-family band of units under 12,000 Btu/h has no room for several heads.

That is an inference from the capacity, not a field in the data, and it is offered as one.

The honest workaround is the one used here — band by capacity, publish the band population and its medians, and let a reader compare an entry against equipment of its own size. It does not recover the single-versus-multi distinction. Nothing available to us does.

Are mini-splits more efficient at every size?

No. The advantage shrinks as units get bigger. Ductless equipment is routinely described as more efficient than ducted equipment, and on the median it is: 22.0 SEER2 against 16.8. HyreHVAC calculation: that gap is not uniform.

Median ductless SEER2 runs 23.5 in the under-one-ton band, 22.5 at one to two tons, 21.5 at two to three, and 20.0 at five tons and above. The advantage narrows steadily as the equipment gets bigger.

HyreHVAC analysis: part of that is physics — a small head serving one room has a short, unrestricted air path and no duct fan to pay for — and part is that the largest ductless outdoor units are, by the nature of the product, driving several indoor units at once.

The data cannot separate those two explanations, for the reason set out above.

What it does establish is that "ductless is more efficient" is a claim about small equipment, and quoting the small-unit numbers while selling a five-ton multi-zone system is not supported by the certified file.

Do larger mini-splits hold more heat in the cold?

Mostly, yes: the second measure inverts the first.

HyreHVAC calculation: median capacity held at 5°F rises across most of the size range — 78% in the smallest band, 86% at three to four tons, 87% at four to five — before falling back to 71% in the small five-ton-and-above group, which holds only 36 families and should not carry much weight.

So the two headline numbers a buyer is shown point in opposite directions across the product range. A small head has the better cooling-season rating; a larger unit tends to hold more of its heating output on the coldest day.

A comparison that reports only SEER2 will systematically favor the smaller equipment, which is one reason ductless rankings elsewhere are dominated by 6,000 and 9,000 Btu/h wall units that could not heat a house.

Ductless equipment is also more likely to carry the Cold Climate designation than the market as a whole — 80.5% of ductless families against 74.0% of all certified families. What that designation certifies, and why the numbers alone do not earn it, is covered on our cold-climate page.