EU heat pumps in 2026: the quietest models are not the most efficient

The EPREL split: quietest models versus most efficient models

The surprise in EPREL’s 2026 heat-pump dataset is that the market’s average outdoor noise is 61.3 dB while average SCOP is 4.55, yet the very quietest and very most efficient models sit in different corners of the table rather than on top of each other (market_index_snapshot). Across 60,989 listed models from 777 manufacturers, that gap is large enough to show up both in the all-market averages and in the leaderboards for quietest models and top SCOP overall (market_index_snapshot).

The quietest 15 entries in the corpus come in at 1.3 dB average outdoor noise, but their average SCOP is only 5.52 across the 13 models in that set with a non-null SCOP value (top_models). By contrast, the top 15 models by SCOP average 6.94 SCOP and report 0.0 dB average outdoor noise in EPREL (top_models). That headline figure needs caution: the highest-SCOP list is dominated by units with 0 dB entries, which likely reflects listing conventions or missing acoustic reporting as much as real-world silence. The corpus shows the ranking, but not why those values were entered that way (top_models).

Even with that caveat, the broader market still points to a split rather than a single “best at everything” cluster. Buyers browsing the full heat-pump catalog or the market index snapshot are not looking at a market where lower noise automatically tracks higher SCOP (market_index_snapshot).

How big the trade-off is in decibels and SCOP points

Within the quietest 15 models, the average outdoor noise is 1.3 dB and the average SCOP is 5.52 for the models where SCOP is present (top_models). The loudest 15 models are not supplied in the corpus, so the exact decibel gap between the quietest 15 and the loudest 15 cannot be answered from the data provided.

Among the top 15 SCOP models, average outdoor noise is 0.0 dB and average SCOP is 6.94 (top_models). Compared with the quietest 15, that makes the top-SCOP set 1.3 dB quieter on paper and 1.42 SCOP points more efficient (top_models). But again, those 0 dB entries mean the “top-efficient models are louder” part of the question is not supported by this corpus. The dataset instead shows the opposite on paper, which is a strong signal that reported acoustic values at the extreme top of the SCOP ranking are not directly comparable to mainstream outdoor-unit noise figures (top_models).

The concrete examples underline the mismatch. WAMAK, s.r.o. TWW 85 WHR is listed at 1 dB and 6.67 SCOP, while Newntide B.V. NE-F1000HCR5TINVM-USC is listed at 1 dB and 5.17 SCOP (top_models). At the efficiency end, the leading Risch model posts 7.0 SCOP at 0 dB, and several Waterkotte models sit at 6.92–6.97 SCOP, also at 0 dB (top_models). So the cleanest supported takeaway is not “quiet always costs efficiency”, but that the quietest named models are not the same models that lead the SCOP table, and the efficiency difference between those two shortlists is material at about 1.4 SCOP points (top_models).

Which heat-pump types sit on each side of the divide

Type explains more of the split than the all-market average does. Air-water heat pumps, which make up 30,452 listings, average 4.54 SCOP and 59.8 dB outdoor noise (type_efficiency). Ground-water units are a much smaller category at 213 models, but average 4.77 SCOP and 58.8 dB (type_efficiency). Water-water heat pumps are tiny in count at 31 models, yet stand out with 6.15 SCOP and only 42.0 dB average outdoor noise (type_efficiency).

That is the clearest type-level divide in the corpus: water-water units are both markedly more efficient and much quieter than mainstream air-water models, with a gap of 1.61 SCOP points and 17.8 dB versus the air-water average (type_efficiency). Ground-water models also beat air-water on both measures, but by a smaller margin: 0.23 SCOP points and 1.0 dB (type_efficiency).

Air-air models complicate the picture because the corpus gives their average outdoor noise at 64.1 dB across 21,065 models, but does not provide an average SCOP (type_efficiency). So any complete efficiency-versus-noise comparison across all major types is limited by missing air-air SCOP aggregation in the supplied data (type_efficiency).

Brand leaders: volume, efficiency, and where they sit on the noise scale

The market is concentrated. Daikin Europe N.V. alone accounts for 14,668 listings, or 24.05% of the EPREL total, with an average SCOP of 4.44 (brand_share). Mitsubishi Electric Europe B.V. follows at 5,575 models and 9.14% share, averaging 4.51 SCOP (brand_share). JOHNSON CONTROLS HITACHI AIR CONDITIONING EUROPE SAS, SUCURSAL EN ESPAÑA ranks third by volume at 5,207 models and 8.54% share, but averages only 4.18 SCOP (brand_share).

Among the highest-volume names, Bosch Thermotechnik GmbH looks stronger on efficiency at 4.69 SCOP across 3,602 models, ahead of Ariston at 4.66 across 2,618 models and Daikin at 4.44 across 14,668 models (brand_share). But the corpus does not provide brand-level average outdoor noise, so it is not possible to say whether the largest manufacturers sit closer to the quiet end or the noisy end of the market overall.

What the shortlist data does show is that the extreme quiet leaderboard is dominated by small names rather than the volume leaders. WAMAK occupies 11 of the top 15 quietest slots, mostly at 1 dB, while Panasonic appears once at 5 dB with a test-labelled entry (top_models). None of the biggest-volume manufacturers dominate the top-SCOP shortlist either; that list is led by Risch, Master Therm, Hoval, Waterkotte, ProCalor, Walter Bösch, Kermi and NIBE (top_models).

Refrigerant effects: why R290, R32, and others cluster differently

The refrigerant story is thin but still revealing. Across the whole market, R32 accounts for 13,935 models, while R290 accounts for only 537; natural refrigerants overall make up 3.27% of listings (market_index_snapshot). That means any buyer using R290 heat-pump catalog filters is searching a much smaller segment than the R32 catalog (market_index_snapshot).

The quietest 10 R290 models in the corpus are all air-air units, with outdoor noise ranging from 2 dB to 59 dB, and none of them report SCOP in the supplied list (top_models). So the corpus does not support a quantified R290-versus-R32 SCOP comparison, nor does it support the claim that one refrigerant systematically explains the quiet-efficient split by itself.

What it does show is that refrigerant choice explains only part of the story because the best efficiency performers in the supplied top-SCOP list do not expose refrigerant data at all (top_models). For buyers, the practical implication is to treat refrigerant as a secondary filter after type, SCOP and declared sound power, and to cross-check definitions in the refrigerants reference and methodology notes.

What buyers can actually do: shortlist rules for installers and consumers

The safe rule from this dataset is to shortlist by type first, not by marketing claims around “quiet” alone. If the project can support water-water or ground-water, the average efficiency and noise numbers are materially better than for air-water: 6.15 SCOP / 42.0 dB for water-water, 4.77 / 58.8 dB for ground-water, and 4.54 / 59.8 dB for air-water (type_efficiency).

Second, compare model pages rather than assuming the quietest leaderboard overlaps with the efficiency leaderboard. WAMAK, s.r.o. BW 11 EVI is listed at 1 dB and 5.4 SCOP, while WAMAK, s.r.o. TWW 48 EVI reaches 6.8 SCOP at the same 1 dB listing (top_models). That spread inside one ultra-quiet cluster is nearly 1.4 SCOP points by itself (top_models).

Third, treat outlier acoustic figures carefully. The corpus contains 0 dB, 1 dB and 2 dB outdoor-noise entries that are far below the all-market average of 61.3 dB (market_index_snapshot; top_models). Installers and journalists should read those as EPREL records first, and physical products second, until the underlying test basis is checked on the product fiche.

Sources

market_index_snapshot — Househeating Pulse · Market Index v1, computed from EPREL Public API. Snapshot: 2026-06-22.
top_models — EPREL Public API via Househeating Pulse catalog. Snapshot: 2026-06-22.
type_efficiency — EPREL Public API · type aggregation. Snapshot: 2026-06-22.
brand_share — EPREL Public API · brand-share aggregation. Snapshot: 2026-06-22.