How the heat pump data is built

Size groups

By rated cooling capacity; one ton is 12,000 Btu/h.

• Up to 1.5 tons (0–20,999 Btu/h cooling)
• 2 tons (21,000–26,999 Btu/h cooling)
• 2.5–3 tons (27,000–38,999 Btu/h cooling)
• 3.5–4.5 tons (39,000–56,999 Btu/h cooling)
• 5 tons and up (57,000 Btu/h and up cooling)

Only manufacturers with a QMID on file (the IRS Qualified Manufacturer ID needed for the 25C tax credit) are ranked: 28 of the 278 manufacturers in this refrigerant group. The others are still counted and listed, marked as not ranked.

1. A product is one outdoor model of one manufacturer. All AHRI certifications of that outdoor unit (its indoor pairings) belong to the product.
2. A product’s headline numbers come from one representative certification: the one with the highest COP at 5°F (ties: the lowest AHRI reference number). Numbers are never combined from different certifications.
3. A model sold with both ducted and ductless indoor units is rated in each type with its best pairing of that type: it is compared with ducted units through its best ducted pairing and with ductless units through its best ductless pairing, and every row shows the AHRI reference and indoor unit of the pairing it is ranked on. Model counts still count each outdoor model once.
4. Duct type follows two rules. In the rankings each certified pairing is compared in one duct type, and AHRI “mixed” pairings (a multi-zone outdoor unit serving ducted and ductless heads at once) count as ductless, because they are the mini-split platform. The Explore filters ask what a model can be installed as, so “Ducted (incl. mixed)” lists every model with a ducted or a mixed certification and “Ductless (incl. mixed)” every model with a ductless or a mixed one.
5. Identical hardware sold under several brands is one row on every leaderboard, with the other badges listed on it (“also sold as”). Two rows are the same hardware when their ranked pairings have the same duct type and capacity, every rating identical (COP and capacity at 5°F, capacity retention, HSPF2, SEER2, EER2), and there is evidence they share hardware: the same manufacturer behind both brands (for example Daikin, Amana and Goodman), or the same outdoor model number under both. Equal ratings alone never merge two manufacturers. The row shown is the brand of the manufacturer that holds the QMID (then the lowest AHRI reference); the collapsed row takes one position. This is display only: every badge still counts in model counts, in its peer group and toward its own manufacturer’s score.
6. Small units rate higher than large ones, and ductless units higher than ducted ones, so a raw number is only compared with comparable units: its peer group, the same capacity size and duct type (ducted or non-ducted) in this refrigerant group. A peer group needs at least 30 models; a smaller one is merged with the next size of the same duct type.
7. A model’s peer percentile is the share of the other models in its peer group that it beats on a rating, ties counting half: 100 is the best in its group, 50 is typical.
8. Only manufacturers with a QMID on file (the IRS Qualified Manufacturer ID needed for the 25C tax credit) are ranked: leaderboards, manufacturer ranks, records and underdogs. Other manufacturers are still counted and listed, marked as not ranked.
9. Top cold-climate picks (the default board): Qualifies when its capacity retention at 5°F (capacity at 5°F as a share of capacity at 47°F) is at least 70% and Capacity retention at 5°F, COP at 5°F, HSPF2 are all rated. Score = the equal-weight mean of its peer percentiles on Capacity retention at 5°F, COP at 5°F, HSPF2 (share of the other models in its peer group it beats, 0–100), all from one certified pairing: in each duct type a model is sold in, its qualifying pairing of that type with the best score; equal scores (one decimal) go to the higher heating capacity at 5°F. Within one duct type a model is scored on its best-scoring pairing that keeps at least 70% of its heat, even when its highest-COP pairing does not; that pairing’s numbers and AHRI reference are the ones shown. The all-sizes board keeps one model per manufacturer (top 25); each size class lists its top 10.
10. All-around best (the second board): Score = 50% heating, the mean of its peer percentiles on COP at 5°F, Capacity retention at 5°F, HSPF2, plus 50% cooling, the mean of its peer percentiles on SEER2 and EER2 (share of the other models in its peer group it beats, 0–100). All five ratings are required; there is no minimum heat kept at 5°F. In each duct type a model is sold in, it is scored on its pairing of that type with the best score; equal scores (one decimal) go to the higher cold-climate score, then the higher heating capacity at 5°F. Each row also shows its cold-climate verdict. The all-sizes board keeps one model per manufacturer (top 25); each size class lists its top 10.
11. The single-rating leaderboards rank by peer percentile, at most one model per manufacturer (top 25). Each row shows the model’s best certification on that rating, compared with the best certifications of the other models in its peer group; equal percentiles go to the model further above its group’s typical value. Inside one peer group the boards rank by the rating itself (top 10).
12. A manufacturer’s score on a rating is the average peer percentile of its models, so the mix of sizes it makes does not move it: a catalog of typical models scores about 50. A model rated in both duct types counts once in each type in that average. It is ranked on a rating once at least 10 of its models are scored on it (again counting each duct type a model is rated in). Raw medians are shown for context only.
13. The efficiency index averages a heating score (COP at 5°F, capacity retention at 5°F and HSPF2 peer percentiles) and a cooling score (SEER2 and EER2). A model gets an index only when all five are known, so a missing rating never helps.
14. Underdogs are verified manufacturers outside the 5 largest, with fewer than 25 products and at least 5 products scored on the rating, whose score beats the combined score of the 5 largest manufacturers’ models.
15. Records are the single highest certified values across every size, labelled with the peer group they come from; they are facts, not a fair comparison across sizes. “Best in class” lists the highest value inside each peer group.
16. Values outside a plausible range for the metric (for example a COP at 5°F above 4.5) are treated as data errors: they are left out and counted on this page.
17. A certification whose NEEP cold-weather curve is physically inconsistent (its COP at a colder temperature reads higher than at a warmer one, outside the normal defrost-derate dip around 17°F) cannot win a spot on any board; it still shows its real certified numbers everywhere else, marked "not ranked: cold-weather ratings look inconsistent" in Advanced view.
18. A certification no longer listed by AHRI cannot win a spot on any board either, since it cannot be bought; it stays visible in every list and on its product page, marked "not ranked: no longer listed by AHRI" in Advanced view.
19. Product counts describe the certified catalog, not sales or installed market share.

Example: a COP at 5°F requirement of ≥ 1.75 means a certification rated 1.74 does not qualify.

What is the difference between a certification and a model?

A certification is one AHRI reference number: one outdoor unit rated with one indoor unit (and sometimes a furnace). A model is one outdoor unit of one manufacturer, with all its certified indoor pairings. This group has 737,594 certifications and 9,640 models.

Which equipment is included?

Active AHRI-certified split air-source heat pump combinations (system types: Air Source Heat Pump, ASHP, Cold Climate Air Source Heat Pump, ccASHP, Split System Heat Pump, Split-System Heat Pump) certified with R-410A. Certifications AHRI no longer lists are included and flagged, not hidden. AHRI test listings (placeholder brands such as "BRAND1") are left out and counted under exclusions. Equipment certified with R-410A, the refrigerant being phased down under the AIM Act.

What are the refrigerant groups?

Low-GWP (R-454B, R-32): Equipment certified with the low-GWP refrigerants required for new installs since 2025. R-410A: Equipment certified with R-410A, the refrigerant being phased down under the AIM Act.

What does CEE 2025 Tier 1 (Split ASHP) require, and how many models meet it?

The CEE 2025 Tier 1 efficiency level for split air-source heat pumps, used by many utility rebate programs. It has these paths; meeting every criterion of one path is enough. Path A (heating-dominated): SEER2 ≥ 16, EER2 ≥ 9.8, HSPF2 ≥ 8.5, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 60%. Path B (cooling-dominated / dual fuel): SEER2 ≥ 16, EER2 ≥ 11, HSPF2 ≥ 8, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 45%. A model counts when at least one of its certified combinations meets the criteria on its own; values are never combined across combinations. In this group 4,663 of 9,640 models (48.4%) meet it.

What does CEE 2026 Tier 1 (Split ASHP) require, and how many models meet it?

The 2026 revision of CEE Tier 1, with higher low-temperature capacity requirements than 2025. It has these paths; meeting every criterion of one path is enough. Path A (heating-dominated): SEER2 ≥ 16, EER2 ≥ 9.8, HSPF2 ≥ 8.5, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 65%. Path B (cooling-dominated / dual fuel): SEER2 ≥ 16, EER2 ≥ 11, HSPF2 ≥ 8, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 50%. A model counts when at least one of its certified combinations meets the criteria on its own; values are never combined across combinations. In this group 4,390 of 9,640 models (45.5%) meet it.

What does CEE Advanced Tier (Split ASHP) require, and how many models meet it?

CEE’s cold-climate tier: full rated capacity and a COP of at least 2.1 at 5°F. Advanced: HSPF2 ≥ 8.5, COP at 5°F ≥ 2.1, Capacity retention at 5°F ≥ 100%. A model counts when at least one of its certified combinations meets the criteria on its own; values are never combined across combinations. In this group 137 of 9,640 models (1.4%) meet it.

What does Xcel Energy Colorado cold-climate ASHP (2025) require, and how many models meet it?

The efficiency thresholds Xcel Energy Colorado uses for its cold-climate air-source heat pump rebate. It has these paths; meeting every criterion of one path is enough. Ducted (ducted equipment): SEER2 ≥ 15.2, EER2 ≥ 10, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 70%. Non-ducted (non-ducted equipment): SEER2 ≥ 16, EER2 ≥ 9, COP at 5°F ≥ 1.75, Capacity retention at 5°F ≥ 70%. A model counts when at least one of its certified combinations meets the criteria on its own; values are never combined across combinations. Not checked here: Program rules beyond ratings (installer enrollment, sizing, documentation) are not checked here. In this group 3,627 of 9,640 models (37.6%) meet it.

What does DOE Cold Climate Heat Pump Challenge require, and how many models meet it?

Rating thresholds from the DOE Residential Cold Climate Heat Pump Technology Challenge. Challenge specification: HSPF2 ≥ 8.5, COP at 5°F ≥ 2.1, Capacity retention at 5°F ≥ 100%. A model counts when at least one of its certified combinations meets the criteria on its own; values are never combined across combinations. Not checked here: Requirements not published in AHRI ratings (cut-in temperature, GWP, controls) are not checked. In this group 137 of 9,640 models (1.4%) meet it.

What does ENERGY STAR Cold Climate require, and how many models meet it?

Units listed by ENERGY STAR with the cold-climate designation (matched to AHRI references). A model counts when at least one of its AHRI certifications is matched to an ENERGY STAR record that carries the cold-climate designation. In this group 2,434 of 9,640 models (25.2%) meet it.

Why are models compared within their size and duct type?

Smaller units and ductless units post higher efficiency ratings than larger and ducted ones, so ranking raw numbers across sizes mostly ranks which sizes a manufacturer makes. In this refrigerant group the median SEER2 is 20.5 for up to 1.5 tons and 15.2 for 5 tons and up, and the median HSPF2 is 9.1 for up to 1.5 tons and 7.8 for 5 tons and up. Every model is therefore compared only with its peer group: the same capacity size and duct type, with at least 30 models per group (a smaller group is merged with the next size). A manufacturer’s score is the average peer percentile of its models, so a maker of large ducted systems and a maker of small mini-splits are judged on the same footing.

What about an outdoor unit sold with both ducted and ductless indoor units?

A model sold with both ducted and ductless indoor units is rated in each type with its best pairing of that type. For example, one outdoor unit certified with a ducted air handler and with wall-mounted heads is compared with ducted units through its best air-handler pairing and with ductless units through its best wall-mounted pairing, and each row names the indoor unit and AHRI reference it is ranked on. Counts of models still count that outdoor unit once; in this group 1,835 models are rated in both types. Duct type follows two rules. In the rankings each certified pairing is compared in one duct type, and AHRI “mixed” pairings (a multi-zone outdoor unit serving ducted and ductless heads at once) count as ductless, because they are the mini-split platform. The Explore filters ask what a model can be installed as, so “Ducted (incl. mixed)” lists every model with a ducted or a mixed certification and “Ductless (incl. mixed)” every model with a ductless or a mixed one.

Why does a leaderboard row say “also sold as”?

Identical hardware sold under several brands is one row on every leaderboard, with the other badges listed on it (“also sold as”). Two rows are the same hardware when their ranked pairings have the same duct type and capacity, every rating identical (COP and capacity at 5°F, capacity retention, HSPF2, SEER2, EER2), and there is evidence they share hardware: the same manufacturer behind both brands (for example Daikin, Amana and Goodman), or the same outdoor model number under both. Equal ratings alone never merge two manufacturers. The row shown is the brand of the manufacturer that holds the QMID (then the lowest AHRI reference); the collapsed row takes one position. This is display only: every badge still counts in model counts, in its peer group and toward its own manufacturer’s score.

What makes a top cold-climate pick, and why not just the highest COP at 5°F?

COP at 5°F is measured at whatever heat output the unit delivers at 5°F. A unit can post a high COP while putting out little heat there, and a unit that loses much of its output in deep cold needs backup heat sooner, however efficient that smaller output is. So a top pick must first keep most of its heat: Qualifies when its capacity retention at 5°F (capacity at 5°F as a share of capacity at 47°F) is at least 70% and Capacity retention at 5°F, COP at 5°F, HSPF2 are all rated. Score = the equal-weight mean of its peer percentiles on Capacity retention at 5°F, COP at 5°F, HSPF2 (share of the other models in its peer group it beats, 0–100), all from one certified pairing: in each duct type a model is sold in, its qualifying pairing of that type with the best score; equal scores (one decimal) go to the higher heating capacity at 5°F. A model sold with both ducted and ductless indoor units is judged in each type on its own pairings, and within a type on its best-scoring pairing that keeps at least 70% of its heat, even when its highest-COP pairing does not. In this group 4,074 of 9,640 models qualify in at least one duct type; 3,641 keep less than 70% of their heat at 5°F and 1,888 have no 5°F capacity rating. The score ranks certified test ratings; it is not a recommendation for a particular home.

What is the all-around best, and how is it different from the cold-climate picks?

The cold-climate picks ask whether a unit holds up in deep cold. The all-around best asks whether it is good at everything: balances heating (5°F output and efficiency, HSPF2) and cooling (SEER2, EER2), each rating against units their size. Heating counts for 50% and cooling for 50%, and a unit missing any of the five ratings is left out rather than guessed. Precisely: Score = 50% heating, the mean of its peer percentiles on COP at 5°F, Capacity retention at 5°F, HSPF2, plus 50% cooling, the mean of its peer percentiles on SEER2 and EER2 (share of the other models in its peer group it beats, 0–100). All five ratings are required; there is no minimum heat kept at 5°F. In each duct type a model is sold in, it is scored on its pairing of that type with the best score; equal scores (one decimal) go to the higher cold-climate score, then the higher heating capacity at 5°F. In this group 7,691 of 9,640 models have all five ratings in at least one duct type.

What is a QMID?

A QMID (Qualified Manufacturer Identification Number) is a short code the IRS gives a manufacturer that has registered as a qualified manufacturer. For heat pumps placed in service in 2025, the federal Energy Efficient Home Improvement Credit (25C), claimed on IRS Form 5695, needs equipment from a qualified manufacturer, identified by a product number that starts with its QMID. Congress ended that credit for equipment placed in service after December 31, 2025. Here a QMID on file is also what lets a brand be ranked against other brands. It says nothing about how well a model performs. Look up a manufacturer’s QMID

Which manufacturers are ranked?

Only manufacturers with a QMID on file: the IRS Qualified Manufacturer ID that 25C tax credit claims need. In this group 28 of 278 manufacturers have one. The others are still counted, listed and given their own pages, marked as not ranked.

How are leaders and ranks decided?

A product is one outdoor model of one manufacturer. All AHRI certifications of that outdoor unit (its indoor pairings) belong to the product. A product’s headline numbers come from one representative certification: the one with the highest COP at 5°F (ties: the lowest AHRI reference number). Numbers are never combined from different certifications. A model sold with both ducted and ductless indoor units is rated in each type with its best pairing of that type: it is compared with ducted units through its best ducted pairing and with ductless units through its best ductless pairing, and every row shows the AHRI reference and indoor unit of the pairing it is ranked on. Model counts still count each outdoor model once. Duct type follows two rules. In the rankings each certified pairing is compared in one duct type, and AHRI “mixed” pairings (a multi-zone outdoor unit serving ducted and ductless heads at once) count as ductless, because they are the mini-split platform. The Explore filters ask what a model can be installed as, so “Ducted (incl. mixed)” lists every model with a ducted or a mixed certification and “Ductless (incl. mixed)” every model with a ductless or a mixed one. Identical hardware sold under several brands is one row on every leaderboard, with the other badges listed on it (“also sold as”). Two rows are the same hardware when their ranked pairings have the same duct type and capacity, every rating identical (COP and capacity at 5°F, capacity retention, HSPF2, SEER2, EER2), and there is evidence they share hardware: the same manufacturer behind both brands (for example Daikin, Amana and Goodman), or the same outdoor model number under both. Equal ratings alone never merge two manufacturers. The row shown is the brand of the manufacturer that holds the QMID (then the lowest AHRI reference); the collapsed row takes one position. This is display only: every badge still counts in model counts, in its peer group and toward its own manufacturer’s score. Small units rate higher than large ones, and ductless units higher than ducted ones, so a raw number is only compared with comparable units: its peer group, the same capacity size and duct type (ducted or non-ducted) in this refrigerant group. A peer group needs at least 30 models; a smaller one is merged with the next size of the same duct type. A model’s peer percentile is the share of the other models in its peer group that it beats on a rating, ties counting half: 100 is the best in its group, 50 is typical. Only manufacturers with a QMID on file (the IRS Qualified Manufacturer ID needed for the 25C tax credit) are ranked: leaderboards, manufacturer ranks, records and underdogs. Other manufacturers are still counted and listed, marked as not ranked. Top cold-climate picks (the default board): Qualifies when its capacity retention at 5°F (capacity at 5°F as a share of capacity at 47°F) is at least 70% and Capacity retention at 5°F, COP at 5°F, HSPF2 are all rated. Score = the equal-weight mean of its peer percentiles on Capacity retention at 5°F, COP at 5°F, HSPF2 (share of the other models in its peer group it beats, 0–100), all from one certified pairing: in each duct type a model is sold in, its qualifying pairing of that type with the best score; equal scores (one decimal) go to the higher heating capacity at 5°F. Within one duct type a model is scored on its best-scoring pairing that keeps at least 70% of its heat, even when its highest-COP pairing does not; that pairing’s numbers and AHRI reference are the ones shown. The all-sizes board keeps one model per manufacturer (top 25); each size class lists its top 10. All-around best (the second board): Score = 50% heating, the mean of its peer percentiles on COP at 5°F, Capacity retention at 5°F, HSPF2, plus 50% cooling, the mean of its peer percentiles on SEER2 and EER2 (share of the other models in its peer group it beats, 0–100). All five ratings are required; there is no minimum heat kept at 5°F. In each duct type a model is sold in, it is scored on its pairing of that type with the best score; equal scores (one decimal) go to the higher cold-climate score, then the higher heating capacity at 5°F. Each row also shows its cold-climate verdict. The all-sizes board keeps one model per manufacturer (top 25); each size class lists its top 10. The single-rating leaderboards rank by peer percentile, at most one model per manufacturer (top 25). Each row shows the model’s best certification on that rating, compared with the best certifications of the other models in its peer group; equal percentiles go to the model further above its group’s typical value. Inside one peer group the boards rank by the rating itself (top 10). A manufacturer’s score on a rating is the average peer percentile of its models, so the mix of sizes it makes does not move it: a catalog of typical models scores about 50. A model rated in both duct types counts once in each type in that average. It is ranked on a rating once at least 10 of its models are scored on it (again counting each duct type a model is rated in). Raw medians are shown for context only. The efficiency index averages a heating score (COP at 5°F, capacity retention at 5°F and HSPF2 peer percentiles) and a cooling score (SEER2 and EER2). A model gets an index only when all five are known, so a missing rating never helps. Underdogs are verified manufacturers outside the 5 largest, with fewer than 25 products and at least 5 products scored on the rating, whose score beats the combined score of the 5 largest manufacturers’ models. Records are the single highest certified values across every size, labelled with the peer group they come from; they are facts, not a fair comparison across sizes. “Best in class” lists the highest value inside each peer group. Values outside a plausible range for the metric (for example a COP at 5°F above 4.5) are treated as data errors: they are left out and counted on this page. A certification whose NEEP cold-weather curve is physically inconsistent (its COP at a colder temperature reads higher than at a warmer one, outside the normal defrost-derate dip around 17°F) cannot win a spot on any board; it still shows its real certified numbers everywhere else, marked "not ranked: cold-weather ratings look inconsistent" in Advanced view. A certification no longer listed by AHRI cannot win a spot on any board either, since it cannot be bought; it stays visible in every list and on its product page, marked "not ranked: no longer listed by AHRI" in Advanced view. Product counts describe the certified catalog, not sales or installed market share.

Are models AHRI no longer lists included?

Yes. Certifications AHRI no longer lists stay in every count and are marked as delisted; this group has 0 of them.

How current is the data?

AHRI Directory of Certified Product Performance: last synced 2026-10-09 (690,993 certifications in this group carry its data). NEEP Cold Climate Air Source Heat Pump List: last synced 2026-10-01 (122,609 certifications in this group carry its data). ENERGY STAR certified heat pumps: last synced 2026-10-08 (47,784 certifications in this group carry its data).

Does a higher model count mean more sales?

No. Counts describe how many models a manufacturer has certified, not how many units it sells or installs.