5G coverage at your address: the map proves nothing

No coverage map proves anything at your address. Atlases are modelled from data declared by the operators, computed on a network carrying no traffic and published twice a year. The only proof lies in a measurement taken on site, and it costs under €15.

That is an uncomfortable position to hold on a site that spends its time quoting the BIPT, the Belgian telecoms regulator. It is nonetheless written in black and white by the regulator itself, in the documentation of its own atlas.

Why does a coverage map prove nothing at your address?

Because a coverage map is not a measurement, it is a calculation. The mobile atlas of the BIPT is built from the coverage data each operator submits, verified through random field checks, then compiled under the most favourable conditions: no network load, theoretical propagation, no competing user on the cell.

The map describes an empty network.

Your evening runs on a full one. At 8 p.m., the cell serving your neighbourhood splits its capacity across every handset attached to it, and 5G loses exactly where it was expected to win, which is in density. Add the publication rhythm, twice a year, and you get a map that can be up to six months behind an antenna switched on or off last week.

Operator maps inherit the same limits, with one extra bias: they are produced by the party selling you the plan. None of them really lies. All of them describe a world tidier than yours.

The atlas has four signal levels, and a 26-point gap

The atlas does not say "covered" or "not covered". It shows four levels, and the level you read changes the answer entirely.

Between the first usable level and the last, the gap is 26 points of households. Same map, same January 2026 dataset, published on 7 May 2026. An operator communicating "over 95 % 5G coverage" and a reader understanding "5G works in my living room" are both talking about this table, and they are not talking about the same row.

DIGI deserves a separate mention, in the other direction. Its own network covered around 34.4 % of the population at the last reading, and its map does not appear in the regulator's atlas, the remainder running on national roaming. Its coverage is therefore not verifiable with the BIPT, only with DIGI.

What does your phone actually measure?

A value in dBm, and it is the only honest figure available to you. It is always negative, and the closer to zero the better the signal: at -75 dBm you are comfortable, at -105 dBm you are at the edge of range, and the antenna bars drawn by the operating system follow no standard scale across manufacturers.

On iPhone, dial *3001#12345#* then the call key: Field Test mode opens. On Android, *#*#4636#*#* opens a diagnostic menu, with wide variation by manufacturer, model and system version. Consumer measurement apps such as OpenSignal or Network Signal Info do the same job with a more readable interface.

A second trap is specific to Belgian 5G. Deployment is overwhelmingly non-standalone, meaning a 5G radio layer riding on the 4G core network. The 5G icon can therefore show while actual throughput stays at 4G level. A speed test on site settles the question in thirty seconds; the icon settles nothing at all.

Six readings to take, and in this order:

• The idle dBm, in the room where you use your phone most, not by the front door.
• The dBm behind a load-bearing wall or in the basement, where signal drops fastest.
• A download and upload speed test at 8 p.m. on a weekday, not on a Sunday morning.
• The same test the next day, at the same hour, to rule out a one-off incident.
• A two-minute call to a landline, to check voice and not only data.
• The home-to-work journey, by train or car, if you spend more than an hour a day on it.

The serious objection: a map eliminates, it does not confirm

An attentive reader will object that if maps prove nothing, they may as well be ignored. The objection is a good one and the answer is no.

A map carries very strong negative value. If the atlas shows "no signal", or only "sufficient signal", at your address for a given network, the matter is closed: that network will not carry indoors at your place, and no test will conjure up an antenna that does not exist. In that direction the regulator is reliable, because an operator has no interest in under-declaring its coverage.

The reverse direction is the one that breaks. "Very good signal" on the map does not guarantee a very good signal at your place, because the model ignores your high-performance double glazing, your metal cladding, your flat at the back of a courtyard and the fifteen neighbours hanging off the same cell. The map rules candidates out. It confirms none.

The method therefore comes in two moves: the map to narrow three networks down to two, the measurement to choose between the two that remain. Our article on 5G coverage municipality by municipality details how to read the atlas in detailed mode, level by level.

How much does a real test of all three networks cost?

Far less than a single month of subscription. A prepaid card gives access to the host network exactly as a plan does, without commitment and without paperwork, and it is the only instrument that reproduces your real conditions. In Belgium the market runs on three mobile networks, Proximus, Orange and Telenet, plus DIGI still building its own; every other brand is an MVNO, a virtual operator renting one of those networks.

Recorded on 31 August 2026, purchase price of the cards at Belgian points of sale:

The calculation that matters: covering all three Belgian networks comes to €15 up front, taking the Telenet card at €0, the Proximus card at €5 and the Orange Tempo card at €10, of which €15 comes back as usable call credit. Against that, the most expensive plan we recalculated on 30 August 2026 costs €1,109.76 over twenty-four months, or €46.24 a month levelled. Fifteen euros to avoid picking the wrong network for two years: the ratio is hard to beat.

One point of honesty, because it is structural. We do not subscribe to these plans in the field and we do not claim to: we compare public price lists and published measurements. The measurement discussed here is yours, at your address, and nobody can take it for you.

Six months, the second test Belgian law hands you

Since July 2022, a Belgian mobile contract can be cancelled free of charge after six months, keeping your number through Easy Switch, with the switch handled by the new operator.

In other words, even signing without having measured anything, your maximum exposure is six monthly payments. It is the most useful safety net in Belgian telecom law, and plenty of subscribers still do not know it exists.

Network first, label second

Once the network is confirmed by measurement, price decides on its own, and the gap is considerable for an identical signal. On the price lists recorded on 30 August 2026, the Mega entry plan comes to €7 at the thirteenth month on the Orange network, yoin to €8.50 at the thirteenth month on the Proximus network, and DIGI shows €9 with no promotion on its own network topped up by roaming.

Three networks, fifteen labels. Measurement tells you which of the three carries at your place; price is read in our 5G plans ranking, sorted on the thirteenth-month price and never on the entry rate. To weigh criterion by criterion, the comparison table lines the eight offers up side by side, and the real 24-month cost gives the total the monthly rate hides.

A map stays a map. Your living room does not model.

Sources: BIPT, network coverage and mobile atlas on the BIPT data portal, January 2026 data published on 7 May 2026; interactive mobile network coverage map, Test Achats; official Orange Belgium prepaid card page and Proximus coverage map, consulted on 31 August 2026; CallMePower and Selectra price records of 30 and 31 August 2026. No link on this page is monetised.