Best Solar Security Lights 2026: 10 Ranked, and Why 3500 Lumens Lasts 24 Minutes

This is the one solar security light on the page where you can check the promise against the hardware and watch it close. Philips publishes an 1,800mAh battery, which stores 6.66 watt-hours, and 1,200 lumens across 302 LEDs. Then it does the thing nobody else does: it states the brightness of each mode as a percentage. Motion-triggered full output lasts 25 seconds. The all-night mode runs at 10%, which is roughly 0.8 watts, and 6.66 watt-hours divided by 0.8 watts is 8.3 hours — precisely the "8 to 10 hours" the bullet claims. A third mode idles at 3% and jumps to full on motion.

Nothing else is remarkable, and that is the point. The sensor reaches 15 metres across a 180 degree arc, which is the longest verified range here and more than double what TECKNET and K KASONIC claim in feet. The three heads rotate 130 degrees, the colour temperature is a cold 7000K, and the panel is named honestly as polycrystalline with no conversion percentage attached — while two rivals under £26 claim 27%, which is the laboratory world record for a silicon cell.

Two cautions. At 4.2 stars it has the joint-lowest average in this comparison, and 302 LEDs producing 1,200 lumens works out at 3.97 lumens per chip, the worst ratio on the page — intelamp gets the same 1,200 from 108 LEDs. The chips are small and there are lots of them. Philips also sells this exact light twice: ASIN B0F4KBNYVS carries the same title and the same 718 ratings at £46.99. We have linked the £26.99 listing.

Two specifications separate this outdoor security light from everything below it, and both are about surviving a British winter. The first is the battery chemistry: lithium iron phosphate rather than the generic lithium-ion everyone else uses. LiFePO4 holds less energy per gram but takes several times more charge cycles and, more importantly here, keeps working when the temperature drops below freezing, which is exactly when a garden light in Manchester is asked to do its job. The second is the panel, quoted at 40 square inches — 258 square centimetres, nearly twice the 139.5 square centimetres of the only rival that publishes a panel size at all.

The light output is modest and honest. One thousand two hundred lumens from 108 LED chips is 11.1 lumens per chip, the best ratio on this page and nearly three times what Philips manages, which tells you these are fewer and larger emitters rather than a wall of tiny ones. Dual PIR sensors cover 180 degrees, the three heads pivot vertically and horizontally, and the listing claims three nights of standby in motion mode — a claim that is at least plausible for a light that only fires for 30 seconds at a time.

The imperial panel measurement is the odd note in a British listing, and it is not the only one: this light also has no published battery capacity, so the three-night claim cannot be checked the way GEARLITE's fifteen-hour claim can. At £33.83 it is also the most expensive light in this comparison, which for one unit rather than a two-pack asks a lot. What you are paying for is a cell that will still be charging in January.

Nine hundred and fifty-four ratings at 4.5 stars is the strongest evidence in this comparison, and the hardware backs a good part of it. The IP66 rating is a step above the IP65 that eight of these ten carry, which matters because the difference is between resisting a jet of water and resisting only spray — relevant on a north-facing wall in a Manchester winter. Three-sided illumination across 300 degrees is wider than anything else here, and the 36-month warranty is the longest offered.

The headline is where it comes apart. Four thousand two hundred lumens is the largest claim in this comparison by 20%, and TECKNET publishes no battery capacity anywhere on the listing. Set it against the biggest cell anyone here does publish, GEARLITE's 2,500mAh or 9.25 watt-hours: 4,200 lumens draws roughly 28 watts, so that battery would empty in twenty minutes, against the "10-12 hours" this listing promises. Four hundred and sixteen LEDs producing 4,200 lumens is also 10.1 lumens per chip, two and a half times what Philips gets from the same class of emitter.

Two smaller things. The claimed 27% panel conversion, described as "7% above average", is the laboratory world record for single-junction silicon — on a £24.99 garden light. And the coverage figure runs backwards against everyone else: this is the brightest claim on the page and it covers "376 ft2", which is 34.9 square metres, while K KASONIC claims 150 square metres from 2,500 lumens. The brightest light here claims the smallest patch of ground.

There is one genuinely good idea in this entire category and CLAONER has it: a real-time battery level indicator built into the housing. Every other listing on this page asks you to believe a runtime figure you have no way to test. This one lets you walk out in December, look at the light, and see whether the panel is keeping up. Coloured mode indicators do the same job for the setting — red for sensor, green for dim-to-bright, blue for constant medium — so you know what the light is doing without cycling through modes in the dark.

The specification is where it gets interesting. Two hundred and eighty-six LEDs producing 2,500 lumens is 8.74 lumens per chip. K KASONIC at number six publishes exactly the same pair — 286 LEDs, 2,500LM — for £16.99, which is £13 less. The two carry separate review pools (930 here against 520 there) so they are not the same listing sold twice, but the identical headline figures point at one factory and two badges. Neither publishes a lumen measurement method, and neither publishes what the 2,500 lumens costs in watts.

Three independently adjustable heads and an IP65 rating are standard for the money. What is not standard is that CLAONER makes no runtime claim at all in its bullets — no "12 hours", no "all night". Given what the arithmetic does to those claims elsewhere on this page, the silence is arguably the most honest thing on the listing.

Three pounds thirty-three per light is the lowest cost per fitting in this comparison, and 952 ratings at 4.4 stars is the second deepest evidence on the page. For lighting a fence line, a side passage and a bin store at once — the actual job most people buy these for — six small units beat one bright one, because a solar wall light only ever illuminates what it points at. The 10cm body is deliberately small, and the listing is right that a compact unit with a proportionally large panel charges faster relative to what it stores.

What the listing does not do is publish a single measurable number. There is no lumen figure, no LED count, no battery capacity and no panel size. The only quantity in the whole set of bullets is a 22% conversion rate, which is plausible for a monocrystalline cell and unverifiable here. Where the others give you numbers that fail arithmetic, this one gives you nothing to check.

It also contains the strangest line collected across this category: a promise to "save hundreds of dollars a year on electricity bills". That is dollars, on a British storefront, for a device that draws well under a watt. Even running a full watt for ten hours every night of the year comes to 3.65 kilowatt-hours, which at current UK rates is about ninety pence. The lights are fine; the copy was written for a different country and a different product.

Two things make this worth a place. It undercuts the CLAONER at number four by £13 while publishing an identical headline specification — 286 LEDs, 2,500LM — which is a strong hint that the same factory supplies both. And it publishes its battery: 2,200mAh, a disclosure only three of these ten make. Four adjustable heads and a remote control are more hardware than the price suggests.

Publishing the battery is also what undoes the marketing. Two thousand two hundred milliamp-hours is 8.14 watt-hours. Twenty-five hundred lumens needs roughly 17 watts even from efficient chips, so a full charge sustains full brightness for about twenty-nine minutes. The listing offers a "permanent on" mode "for 4 hours at night", which would take 67 watt-hours — eight times what the cell holds. Run the sum the other way and four hours of light from 8.14 watt-hours means about 2 watts, or roughly 300 lumens. That is a perfectly usable porch light. It is not 2,500 lumens.

The rest is ordinary and mostly reasonable. The panel is described as polycrystalline with a 21% conversion rate, which is above what polycrystalline silicon actually achieves — 21% is monocrystalline territory, and either the panel or the number is mislabelled. Coverage is quoted as "1615 sq ft/150 m2" and detection as "25FT/7.6M", both imperial-first on a British listing, though at least the conversions are correct.

The split panel is the argument for this one and it is a good argument. Every all-in-one solar floodlight has to be mounted where the sun is, which in a British garden is usually not where the intruder is. Separating the panel from the lamp with five metres of waterproof cable lets you put the light over a north-facing back door and the panel on the sunny gable end. In a country where December daylight in Manchester runs to seven and a half hours and most of it is overcast, that flexibility does more for real-world performance than any lumen figure.

The numbers are thin. Two hundred and twenty-six LEDs are advertised and no lumen figure is published anywhere, so the LED count is doing all the selling. Coverage is stated as 30 square metres, the smallest claim in this comparison — which, sitting next to TECKNET's 4,200 lumens over 34.9 square metres, mostly shows how little these area figures mean. The sensor is the weakest here at 10 to 17 feet, which is 3 to 5.2 metres; Philips reaches 15 metres.

At 4.4 stars from 594 ratings the evidence is solid, and a remote control at £19.99 for two units is generous. Note the listing still describes itself as "2023's newly upgraded" and quotes the same 22% conversion rate as three other brands on this page, which suggests the figure comes from a panel supplier's datasheet rather than from anyone's measurement.

Kolpop builds the same sensible thing as MinPea — panel separated from lamp, five metres of detachable cable, remote control, IP65 — and adds a convex lens to concentrate the beam, which the listing says gives ten metres of throw. Concentrating light with optics rather than adding chips is the correct engineering answer in this category, and it is the same approach GEARLITE uses at number ten. The bullets also include a genuine, useful troubleshooting section explaining why the light will not come on in daylight and how to check the connector alignment, which is more customer support than most of these listings attempt.

What is missing is everything you would measure it by. No lumens. No LED count. No battery capacity. No panel size. The only figure offered is a 22% conversion rate, the same number MinPea, vighep and CZHHMART all quote, and a claim of "8-12 hours of continuous bright lighting" attached to an "extra-large capacity battery" whose capacity is never stated. Given that the three listings on this page which do state their battery all fail that arithmetic badly, an unstated capacity behind an all-night claim is not reassuring.

At 4.2 stars from 438 ratings it has the shallowest review pool and joint-lowest average in this group, and at £20.96 it costs a pound more than the MinPea that does the same job with a published LED count and a stated coverage area.

At £9.99 this is by some distance the cheapest way onto this page, and 791 ratings at 4.4 stars is better evidence than seven of the nine lights above it can show. It publishes its battery too — 1,200mAh — which puts it in the honest minority here, and the bullets are unusually specific about installation, recommending a mounting height of 3 to 5 metres and supplying screws, tape and a guide.

That published battery is also the most instructive number in this comparison. Twelve hundred milliamp-hours is 4.44 watt-hours. Spread across the "8-12 hours" of illumination the listing promises, that is 0.44 watts — call it 65 lumens on generous assumptions. Divided among the advertised 248 LEDs, each chip is receiving under two milliwatts. The LEDs are real and the light works; what the arithmetic shows is that 248 of them exist to make a number in the title, not to make light. Every listing on this page is doing the same thing. This is simply the only one cheap enough and open enough that you can prove it.

One oddity worth naming. The first bullet is written in English and then, mid-sentence, switches to Spanish: "Features a Panel solar de 22% de eficiencia y sensor de movimiento para mayor seguridad y ahorro energetico". An untranslated fragment sitting inside the headline feature of a British listing tells you how much attention the specification received before it was published.