What does a heat pump cost to run?

Running cost on a heat pump is one short equation. Annual heat demand, divided by efficiency, multiplied by what you pay per unit of electricity. Add the standing charge and you are done.

Heat demand depends on the building. A leaky 1930s semi-detached uses a lot more than a 2010s flat. Most of UK housing stock falls between 8,000 and 15,000 kilowatt-hours of useful heat per year, which is the building's actual heating output rather than the gas or oil that goes in. Hot water adds another 1,200 to 3,000 kilowatt-hours depending on the household.

Efficiency on a heat pump is its Seasonal Coefficient of Performance, or SCOP. The Energy Saving Trust uses 3.0 as a typical UK figure. A well-designed install on a modest flow temperature reaches 3.5 to 4.0. We come back to where that number actually lands further down.

Tariff is what you pay per kilowatt-hour of electricity. The April 2026 Ofgem price cap sets the average at 24.7 pence. Heat-pump-aware tariffs like Octopus Cosy or Octopus Go have a cheap window and a more expensive day rate. Whether they actually beat the cap depends on whether you use them properly.

The chart below is what those numbers add up to, for a standard UK home using 12,000 kilowatt-hours of useful heat per year. The HeatPass result page does the same math against your specific home and shows the band you should plan around.

The chart below is the year-one heating and hot-water bill for a standard UK home using 12,000 kilowatt-hours of useful heat. Every row uses the same heat demand. The only thing that changes is the fuel and the efficiency of the kit that converts it into warm air and hot water.

Three things to read off the bars. First, gas wins narrowly at today's prices. The Ofgem cap unit rate fell again from April 2026 to 5.7 pence per kilowatt-hour for gas and 24.7 pence for electricity, which puts a typical heat pump install around £300 a year more than a condensing gas boiler on the same demand. That is the honest baseline.

Second, the gap closes when the install does the work. A SCOP of 3.5 instead of 3.0, plus an Octopus Cosy tariff used with some scheduling discipline, pulls a heat pump under the gas line on the standard-home demand we have used here. We unpack what each of those two changes is worth in the next two sections.

Third, the comparison flips hard against oil, LPG and storage heaters. A heat pump beats the post-spike oil figure by around £500 a year and the LPG row by around £600. Against electric storage heaters it is close to a third of the bill. The off-grid case for a heat pump is much stronger than the gas case, even before any comfort improvement is counted.

The pre-spike heating-oil reference shows what the comparison looked like in February 2026, before disruption to shipments through the Strait of Hormuz roughly doubled UK kerosene retail prices. The BoilerJuice national average peaked at 134 pence per litre in March 2026 and has been settling slowly since. Oil pricing is volatile through 2026, and the chart row you should plan around is whichever figure looks most like the price your local supplier quoted last week.

Most homeowners reading the chart above expect a heat pump to be dramatically cheaper than gas. Heat pumps are three times more efficient. The arithmetic should follow. The reason it doesn't is the price ratio between electricity and gas, and the UK has one of the worst in Europe.

The bottom line is two numbers. UK electricity costs about 4.3 times more per unit than gas. A heat pump is only about three times more efficient than a boiler. The gap between those two numbers is why heat pumps don't undercut gas at today's prices. The arithmetic underneath is straightforward. A heat pump's running cost is the heat it delivers, divided by its SCOP, paid for at the electricity rate. A gas boiler's running cost is the heat it delivers, divided by its efficiency, paid for at the gas rate. For both bills to be the same, the ratio of electricity to gas needs to roughly match the ratio of heat-pump SCOP to boiler efficiency.

A typical UK heat pump runs at SCOP 3.0. A typical modern gas boiler runs at 90 per cent efficiency. The ratio between them is about 3.3. So if electricity cost 3.3 times what gas costs, a heat pump would tie a boiler. If electricity cost less than that, the heat pump would win. Above that, gas wins. The UK currently sits around 4.3× under the April 2026 cap, which is why the chart in the previous section has gas just under the heat pump and not far above it.

Most of mainland Europe runs the same arithmetic on a very different ratio. Nesta's research notes France runs at roughly 1.7×, with Germany and most of the rest of the EU between 2 and 3. That is why a heat pump genuinely undercuts gas in those markets, and why heat-pump uptake has been so much faster there than here. The technology is the same. The unit prices are not.

Two policies move this ratio. Removing the social and environmental levies that currently load onto electricity bills shifts the UK ratio down by about half a point on its own. Increasing the share of UK generation that comes from renewables, where marginal cost is essentially zero, shifts it down further over time. Both are explicit in current government direction of travel. Today's ratio is not next year's, and the gap on the chart in By fuel above will close as the ratio does.

The Seasonal Coefficient of Performance is the average ratio of useful heat delivered to electricity used, over a whole heating season. A SCOP of 3.0 means the heat pump produced three kilowatt-hours of heat for every one kilowatt-hour of electricity it drew. SCOP includes the fact that performance drops on cold days, which is why it is the figure that matters for a year-one bill. The instantaneous ratio at any one moment is the Coefficient of Performance, or COP, and it ranges from roughly 4.5 on a mild autumn morning down to 2.5 on a cold January night.

The number you should plan around depends on three things. The flow temperature, meaning how hot the water is when it leaves the heat pump and enters the radiators, that the system is designed for. How well the radiators or underfloor are matched to that flow temperature. And how disciplined the installer was about commissioning. A heat pump asked to run at a 55°C flow temperature into undersized radiators will sit around SCOP 2.6 to 2.8. The same hardware on a 40°C flow temperature into properly sized emitters will reach 3.5 to 4.0. The radiator question is the one we cover in detail in the property-suitability guide.

The Energy Saving Trust's published illustration uses a SCOP of 3.0, and Nesta uses 3.0 in its bill modelling. Both are conservative, intended to describe a typical acceptable install rather than a best case. The opt-in monitored fleet at HeatPumpMonitor.org reports an average closer to 3.9 across several hundred UK systems with continuous monitoring. That figure is biased upward, because the homeowners who fit detailed monitoring also tend to commission carefully and choose installers who design for low flow temperatures. But it is the only public dataset of observed UK reality, and the gap between 3.0 and 3.9 is roughly a £200 a year swing on a typical bill.

Two practical takeaways. First, if your installer's heat-loss calculation puts your system at a flow temperature of 50°C or lower, expect a SCOP of 3.5 or higher. Above 55°C, expect 2.8 to 3.2 and plan around the lower number. Second, if a year into the install your bill is materially higher than the heat pump's quoted SCOP suggests, ask the installer to review the weather compensation curve. Most under-performance traces back to a flow temperature set too high during commissioning, which is fixable in software rather than hardware.

Wall fabric matters too. A poorly insulated wall pushes the design flow temperature up, which drops realised SCOP. The Victorian house guide plots that effect on a solid 9-inch brick terrace. The 1930s house guide plots it on a cavity-walled semi.

Three UK suppliers run tariffs designed around heat pumps and electric vehicles. Each shapes the day around a cheap window plus a standard rate, and Cosy adds a short early-evening peak. The shape and the rates as of May 2026:

Tariff	Cheap	Standard	Peak
Octopus Cosy	14p, 9 hrs/day	32p, 12 hrs/day	50p, 3 hrs/day
Octopus Go	10p, 5 hrs/day	31p, 19 hrs/day
Good Energy Heat Pump	14p, 7 hrs/day	30p, 17 hrs/day

Octopus Cosy

Cheap

14p 9 hrs/day

Standard

32p 12 hrs/day

Peak

50p 3 hrs/day

Octopus Go

Cheap

10p 5 hrs/day

Standard

31p 19 hrs/day

Peak

none

Good Energy Heat Pump

Cheap

14p 7 hrs/day

Standard

30p 17 hrs/day

Peak

none

The thing to understand about all three is that the cheap window only beats the Ofgem cap if you actually shift your consumption into it. A heat pump left on its default schedule, doing roughly time-prorata consumption across the day, lands at almost exactly the cap rate on Cosy. The savings come from pre-heating the home overnight, scheduling the hot-water cylinder to reheat in the cheap window, and avoiding the early-evening peak. With that discipline, Cosy's effective rate falls to around 22 pence per kilowatt-hour, roughly 10 per cent below the cap.

The discipline is harder than it sounds in a typical household. A pre-heat overnight only works if the building holds heat reasonably well. If the home is poorly insulated, the heat leaks away before morning, and you end up paying the cheap rate for heat the fabric throws back out by 9 a.m. The same building that scores well on a heat-pump install (decent walls, decent loft, sensibly sized radiators) also happens to be the one that gets the most out of a time-of-use tariff. The two questions are linked.

Practical guidance. If you are on Cosy or considering it, set the heat pump to pre-heat from 4 to 7 a.m., schedule the cylinder reheat for the same window, and set the weather compensation curve to its lowest workable flow temperature. If you can shift the dryer, dishwasher and any electric vehicle charge into the cheap windows too, the tariff starts to pay for the effort. If you can't, the standard cap is fine and you're not leaving much on the table.

Most of a heating bill is the building. Walls, loft, window area and cold air leaking through draughty doors lose heat at roughly the same rate whether one person or five live behind them. The household-size effect on the running cost is mainly hot water, plus a smaller second-order effect from how often the heating is on.

BRE and Energy Saving Trust both use three brackets for hot-water demand. A household of one or two uses around 1,200 kilowatt-hours a year. Three or four people use around 2,000. Five or more use about 3,000. On a heat pump at SCOP 3.0 that's roughly £100, £165 and £247 of hot-water cost per year at the April 2026 cap. The increment per extra occupant is about £50 to £80 a year on a heat pump, less on gas.

The HeatPass result page asks for household size explicitly because the alternative is to assume the middle bracket and quietly under-cost a five-person home. We model the three buckets directly so the year-one figure is honest about your specific situation rather than an average that fits no-one.

One non-obvious wrinkle: a five-person household with a combi boiler today often runs the boiler much harder than a two-person household, even though they live in the same kind of building. Hot water is drawn more often, the boiler cycles more, and a proportion of the heat output is wasted in cycling losses. The £/year difference between household sizes is therefore larger on a combi boiler than on a heat pump with a hot-water cylinder, which spreads the load across a thermal store. Switching to a heat pump tends to shrink the per-occupant gap, not widen it.

A worked example. Take a 90 square-metre 1980s semi-detached, cavity walls filled, loft insulated, three occupants. That home uses around 10,000 kilowatt-hours of useful space heat per year plus 2,000 of hot water, so 12,000 in total. The heat pump is correctly sized at 7 kilowatts, runs at a 45°C flow temperature into modestly upgraded radiators, and reaches a real-world SCOP of about 3.4.

On Octopus Cosy with sensible scheduling, that translates to roughly 3,500 kilowatt-hours of electricity at an effective rate of 22 pence, plus the standing charge. Year-one bill: about £960. Their previous gas bill at the April 2026 cap on the same demand: about £870. The heat pump costs them about £90 more per year, comfortably below the £150-£200 a year that running the gas boiler one more degree warmer for the same comfort level used to add. The household replaces the gas safety check cost with a heat-pump service, roughly cancelling each other out.

Three things change that picture. A SCOP of 3.0 instead of 3.4 (poorer commissioning, higher flow temperature) adds about £100 to the year. A worse tariff, like staying on the cap rather than moving to Cosy, adds about £150. A better tariff with aggressive scheduling discipline saves about £80. The honest planning range for this household is £900 to £1,200 in year one.

Add solar and a battery to the same install and the number drops further. UK owners on heat-pump-plus-solar-plus-battery setups typically report annual savings of £480 to £700 against their old gas bill, on similar-sized homes. The best cases reach net zero across the year when an electric vehicle is in the mix and a heat-pump-aware tariff keeps the battery working every day. None of that changes the year-one number for a heat-pump-only install, but it is the upgrade path that owners thinking three or four years out tend to land on.

The number behaves like a gas bill once it's in. It moves with the weather year to year, with the tariff each spring and autumn, and with whatever fabric improvements you make later. The biggest lever after the install is the flow temperature on the weather compensation curve, which an installer can re-tune. The second biggest is the tariff review every six months when the cap moves. The cost picture for the install itself, gross and net of the £7,500 grant, is in the cost-and-grants guide.

The year-one running cost is one number, and it is a fair enough way to think about the money side. But owners who have lived with the system for a couple of years usually have two things to add that don't show up in any spreadsheet.

The first one is whether heat pumps should be expected to pay back at all. A replacement boiler doesn't pay back. A new washing machine doesn't either. A heat pump is an appliance, and asking when it pays for itself puts it under a test no other household appliance has to take. The fair comparison is appliance to appliance, gas combi against heat pump, both with their own running costs and their own service intervals.

The second is comfort. Once you stop heating the house in two-hour bursts and start running it weather-compensated all day, the place gets warmer in a way the spreadsheet doesn't see. Owners describe it as the house being a nicer place to be in winter. The numbers compare gas-at-intermittent-heat against heat-pump-at-constant-heat, but those are not the same comfort level. The heat-pump house is warmer.

The year-one number still matters. It just isn't the whole picture.