Heat pump running costs versus a gas boiler, assumptions shown

The short answer

Three efficiency figures get used interchangeably and they are not the same thing. COP is the instantaneous ratio of heat delivered to electricity consumed, and it moves constantly with outdoor temperature and flow temperature, so a single COP tells you almost nothing about a year of bills. SCOP is a seasonal figure from a standard test, always quoted at a stated flow temperature, which is why datasheets say things like SCOP 4.84 at 35C. SPF is what the system actually delivered in your house across a real year. Only SPF pays the bill, and it usually sits below the datasheet SCOP.

Four assumptions decide the whole answer

• Seasonal efficiency. The government funded Electrification of Heat project monitored 742 installations and recorded a median in situ SPF of 2.80 for air source systems. A well designed modern installation reaches 3.5 to 4.0. Hybrid systems that keep the gas boiler alongside did worse, at a median heat pump efficiency of 2.37.
• Flow temperature. Heat pumps run most efficiently at 35 to 45C flow, against 70 to 80C for a conventional boiler system. MCS design for radiator retrofits commonly lands at 45 to 55C, and every degree above the minimum costs efficiency.
• Tariff. The comparison below uses capped standard rates. A heat pump tariff with cheap windows changes the outcome, and so does sitting on a flat rate and never shifting any load.
• Insulation and heat demand. How many kWh of heat the house needs in a year, which comes out of a room by room heat loss calculation rather than a rule of thumb, and which sets the flow temperature your existing emitters can get away with.

The worked example, 10,000 kWh of heat a year

Ofgem price cap averages for 1 July to 30 September 2026, Direct Debit, VAT included. Rates vary by region and payment method. The cap for 1 October to 31 December 2026 is published by 26 August 2026, so treat every figure on this page as quarter dated.

Take a house that needs 10,000 kWh of useful heat a year. That number is an assumption for the sake of the arithmetic. In a real quote it comes from the heat loss survey, not from an average. Assume the gas boiler runs at 90 percent seasonal efficiency, the standard modelling figure and a generous one for an older system, so it burns 11,111 kWh of gas to deliver that heat.

Fuel cost only, standing charges excluded. Our arithmetic, worked from the capped unit rates in the table above.

Read that honestly. A median performing heat pump on standard capped rates costs roughly £119 a year more than the gas boiler. A good one saves £68. A very good one saves £161. The design decides which row you land in, not the badge on the outdoor unit, and the trial data shows the spread between installations is wide enough to swallow the entire argument in either direction.

Flow temperature and insulation are the same lever

Efficiency falls as flow temperature rises, and it falls hard. Samsung publishes a SCOP of up to 4.84 at 35C flow on its R290 monobloc range, and 2.82 to 3.03 at 65C flow across its high temperature range. Same manufacturer, same technology, and the gap between those two numbers is the difference between a heat pump that beats gas comfortably and one that does not beat it at all. A system asked to produce 65C water because nothing else in the house changed is a system that will disappoint you on the bill.

Insulation belongs in this conversation, though not the way it is usually framed. There is no insulation condition and no EPC requirement on the Boiler Upgrade Scheme eligibility page, so nobody is going to block the installation. Fabric matters for arithmetic reasons: less heat loss per room lets the designer specify a lower design flow temperature using the emitters already on the wall, and a lower flow temperature raises the SPF, which is the number this entire comparison turns on. The Electrification of Heat project installed across Victorian terraces, 1960s flats and pre war semis and concluded no property type or era is unsuitable. What varied was installation quality, not the age of the building.

That is also why radiators surface in a running cost conversation. Radiator outputs on a manufacturer chart are rated at Delta T 50, meaning 75C flow, 65C return and a 20C room. At 45C flow and 40C return the same radiator delivers roughly 42 percent of its rated output, so a panel rated at 1,500 W gives you about 630 W. Holding the same room temperature at heat pump flow temperatures needs roughly 2.4 to 2.5 times the Delta T 50 rated output. That does not mean every radiator gets replaced, because plenty of British systems are already oversized, and which rooms need bigger emitters is decided room by room in the heat loss calculation. Our heat pump guidance covers the design side in more detail.

A heat pump tariff changes the answer

• Cosy Octopus. Three cheap windows a day, 04:00 to 07:00, 13:00 to 16:00 and 22:00 to 00:00, priced 51 percent below the regional day rate, with a peak period from 16:00 to 19:00 at 50 percent above it. Open to any make of heat pump, electric boiler or electric radiators, on a SMETS2 or compatible SMETS1 meter.
• E.ON Next Pumped. Super off peak from 22:00 to 06:00, off peak from 06:00 to 16:00 and 19:00 to 22:00, peak 16:00 to 19:00, on a 12 month fix with no exit fees. The per kWh rates are quoted per property rather than published.
• British Gas Heat Power. Half price electricity from 13:00 to 16:00 and 00:00 to 07:00, 70 hours a week, applied as a weekly credit on the bill rather than as a different unit rate. Any make of heat pump or electric heating system, credit customers only.
• Good Energy publishes a heat pump rate of 14p per kWh across seven off peak hours, 05:00 to 09:00 and 13:00 to 16:00, for its own supply customers.
• OVO Heat Pump Plus, which delivered an effective 15p per kWh through a monthly rebate, was withdrawn from 1 February 2026 and is closed to new customers. Tariffs move, and a running cost promise built on one that has gone is worth nothing.

Illustrate what that does to the sums. Take the median SPF 2.8 case again and assume load shifting brings the blended rate down to 20p per kWh, which is an assumption on our part rather than a quoted rate. The same 3,571 kWh then costs £714 a year and undercuts the £814 gas boiler even at trial median efficiency. That is the decisive move in the whole comparison: an average heat pump on a good tariff beats a gas boiler, while an excellent heat pump on a flat capped rate can still lose to one. Rates reset every quarter, so check the dated picture in our tariff guidance before building a decision on any single number.

Work out your own number