Live speed test · free & no app
Test your real internet speed right now
One click measures your download, upload, ping and jitter against Cloudflare's global test servers, live in your browser. Remember what this page is about: it reads the path from this device, across your wifi, to one server, at this moment, so run it wired and off-peak before you judge your line.
What this actually told you: your download came in at with a ping. That is the speed along this path right now, including your wifi and device, not a verdict on your NBN plan. See exactly what your number means ↓
The test downloads and uploads a few megabytes of data to a Cloudflare server to measure throughput, so it uses a small amount of your data allowance. Results vary with wifi, your device and the time of day. For the truest reading, plug in by ethernet and test off-peak.
The short answer
What an internet speed test actually measures
An internet speed test is one of the most useful buttons on the web and one of the most misread. It does not measure your NBN plan. It measures the whole path from your device, across your wifi, out to a single test server, at a single instant in time. That distinction is the difference between fixing your internet and wasting money on it, because a slow result is far more often your wifi, your device or peak-hour congestion than a problem with the line itself. This page explains exactly what a speed test does and does not measure, why a single reading is the wrong number for choosing a plan, and how to run the test properly so it tells you something true.
Here is the core takeaway in one line: a speed test measures the slowest link in the chain from your device to one server at one moment, which is usually your wifi or the peak hour, not your NBN line, so a single reading is the wrong number for deciding whether to upgrade your plan. That reframe fixes the two costly mistakes people make: they run the test on wifi at 8pm, see a low number and blame the line; and they react by buying a bigger speed tier, which does nothing for a wifi or congestion problem. To test the line, plug in by ethernet and test off-peak. To choose a plan, ignore the one-off test entirely and compare the Typical Evening Speed.
Reframe the assumption: the speed test is a stethoscope, not a verdict. It tells you the speed along one path right now, including your wifi, your device and the current congestion. Change any of those and the number changes, which is exactly why a healthy line can post a poor result, and why a bigger plan rarely fixes it.
The blind spot
Why most speed-test guides lead you astray
Most "how to test your internet speed" articles do one of three unhelpful things, and each one nudges you toward spending money on the wrong fix.
First, they treat the result as a verdict on your plan. They tell you to hit the button, read the download number and compare it to your tier, as if the figure came straight off the line. It did not. It came off your wifi, your device and the route to one server, so a low reading gets blamed on the NBN when the bottleneck is sitting in your loungeroom.
Second, they ignore when and how you test. A test run over wifi at 8pm is measuring two of the biggest variables at once: your wireless signal and peak-hour congestion. Guides that do not tell you to plug in by ethernet and test off-peak are handing you a number that cannot isolate the line, then asking you to make a decision with it.
Third, they jump straight to "upgrade your plan". It is the most expensive advice on the page and usually the least relevant. A bigger speed tier does nothing for a wifi dead zone, an ageing laptop or a retailer that skimps on evening bandwidth. The result mis-diagnoses the problem, and the recommended cure treats the wrong organ.
What a speed test really sees, piece by piece
Expert analysis: the chain a speed test actually measures
A test reads the slowest link, and it is rarely the line
A speed test sends data between your device and a test server and times it. What it reports is the speed of the slowest link in that whole chain: your device's network card, your wifi, your home cabling, your router, the NBN line, your retailer's network and the route to the server. Any one of those can be the bottleneck, and in a typical home the weakest link is the wifi, not the fibre. This is why the same line can post 90 Mbps wired and 30 Mbps from the back bedroom: the line never changed, the wifi did.
Wifi and the device cap the result before the line ever does
Wireless is shared, contended and easily blocked. Distance, brick walls, a microwave, a neighbour's network on the same channel and a crowd of connected devices all eat into a wifi signal, so the figure you see can sit well below what the line could carry. The device matters too: an old phone or a laptop with a tired wifi chip simply cannot pull a high speed regardless of the plan. None of this is an NBN problem, and none of it is fixed by buying a faster tier. It is fixed by a wired connection, better router placement or a mesh node.
Time of day measures congestion, not capacity
The NBN is a wholesale network run by NBN Co, which sells your retailer (an RSP, or Retail Service Provider, such as Telstra, Optus, Aussie Broadband or TPG) a shared pool of bandwidth. The retailer decides how much of that pool to buy per customer. In the quiet daytime there is plenty to go around, so a midday test looks great; in the 7pm to 11pm peak, the whole street is online and a retailer that bought too little lets everyone slow down. A speed test at 8pm is therefore measuring the retailer's bandwidth decision as much as anything physical, which is precisely the variable a single off-peak test cannot see.
The honest plan number is the Typical Evening Speed, not your test
This is why a one-off test is the wrong tool for choosing a plan. The number that predicts your real experience is the Typical Evening Speed (TES): the download speed a retailer is required to publish for the busy 7pm to 11pm window. It already bakes in the congestion your single test cannot capture. The Australian Competition and Consumer Commission (the ACCC) reinforces this with its Measuring Broadband Australia programme, which tests real homes and reports which retailers actually hold their speed at peak. Use your own test to diagnose your gear; use the TES and the ACCC results to choose a plan.
Interactive explainer
Interpret your speed test result
Enter your result, your plan tier, whether you tested on wifi or wired and the time of day. It tells you what the number actually means, wifi-limited, line-limited, peak congestion or a healthy line, and the right next step. It does the diagnosis the raw figure cannot.
What does my speed test result mean?
Indicative diagnosis only. Always retest wired and off-peak before changing plan.
How and when did you test?
What this means
How it reads your input: a result is treated as healthy when it reaches roughly 80% or more of the tier on a wired off-peak test (the share the line should deliver). A wifi result is judged loosely because wifi naturally loses speed; a low wired off-peak result points at the line or connection type; a low result only at peak points at congestion. Sources: ACCC Measuring Broadband Australia, NBN Co, Canstar Blue 2026. Always retest before acting.
| Test conditions | What it means | The right next step |
|---|---|---|
| Wired, off-peak, result near your tier | Your line is healthy | Nothing to fix on the line. If wifi feels slow, the problem is between the router and your device. |
| Wifi result far below the wired result | Wifi-limited, not line-limited | Move closer to the router, change the channel, or add a mesh node. Do not upgrade your plan. |
| Wired result fine by day, poor at 7 to 11pm | Peak-hour congestion | Compare retailers by Typical Evening Speed; the line is fine, the retailer skimped on bandwidth. |
| Wired, off-peak, result well below your tier | Possible line or connection-type limit | Likely FTTN copper capping the tier, or a line fault. Check your connection type before paying for more. |
| Activity | Speed needed | What actually matters |
|---|---|---|
| General browsing, email, social media | 1 to 5 Mbps | Light, bursty; latency matters more than raw speed. |
| HD video streaming (per stream) | ~5 to 7 Mbps | A single HD stream barely dents a modern plan. |
| 4K / Ultra HD streaming (per stream) | ~25 Mbps | The single biggest household demand; multiply by simultaneous streams. |
| Video calls / working from home | ~3 to 5 Mbps | Needs steady upload and low latency, not big download numbers. |
| Online gaming | ~3 to 15 Mbps | Tiny data, but punishes high latency; aim for ping under 50 ms. |
| Large uploads or cloud backups | 50+ Mbps | The one activity that genuinely rewards a top tier. |
What this looks like in real homes
How a misread speed test costs Australians money
The mechanics above are not theory. They are exactly how households end up paying for the wrong fix:
The upgrade that fixed nothing
A family runs a wifi test from the kitchen, sees 28 Mbps on their NBN 100 plan and upgrades to NBN 250 to "speed things up". The line was always healthy; the kitchen was just a wifi dead zone. They now pay top-tier prices and still get 28 Mbps in the kitchen, because the bottleneck was never the line.
The 8pm panic test
A renter tests at 8pm on a Sunday, sees the number tank and assumes the connection is broken. A retest at 11am the next day comes back near the tier. The line was fine all along; the Sunday-night result was peak-hour congestion, a retailer issue best solved by comparing Typical Evening Speeds, not by ringing tech support.
The gamer chasing the wrong number
A gamer fixates on getting a bigger download figure and upgrades to NBN 1000, but the lag does not budge. Their real enemy was latency, not throughput. A clean wired connection and a retailer with low ping would have helped; a fatter download pipe did not, because gaming uses tiny data and punishes high ping.
The slow laptop blamed on the NBN
A household tests on an ageing laptop with a tired wifi chip, gets a poor result and considers switching providers. The same line tested on a new phone, wired, posts near the tier. The device was the cap, and no plan change or provider switch would have lifted it.
The insider insight
A speed test diagnoses your home, the TES predicts your plan
Here is the part most guides never say plainly. There are two completely different questions a speed test gets asked to answer, and it can only honestly answer one of them. "Is my gear working right now?" is a diagnostic question, and a wired off-peak test answers it well. "Which plan should I buy?" is a forecasting question, and a one-off test is useless for it, because the plan's real-world speed is the evening congestion your daytime test cannot see. Confusing the two is how people end up upgrading a plan to cure a wifi problem, or staying on a congesting retailer because their lunchtime test looked fine.
The non-obvious truth: your speed test is for diagnosing your own equipment, and the Typical Evening Speed is for predicting your plan. Run the test wired and off-peak to check the line and your gear; then ignore it entirely when choosing a plan and compare retailers on their published evening speed instead. The households that get value test to diagnose, not to decide, because the number that decides is the one the retailer publishes for 8pm, not the one you catch at noon.
The practical consequence: never let a single speed test, especially a wifi one, push you into a bigger plan. Diagnose first, and let the Typical Evening Speed do the deciding.
Grounded in the analysis
What you should actually do with a speed test
Specific moves that follow from what a test really measures, not generic advice.
Test the line, not your loungeroom
Plug your device into the router with an ethernet cable, close other apps and get the household offline, then test off-peak in the middle of the day. Only a wired off-peak test isolates the line from your wifi and the evening crowd.
Diagnose before you spend
Compare the wired off-peak result to your tier using the interpreter above. If only wifi or only the evening is slow, the fix is router placement, a mesh node or a wired link, not a bigger plan. Upgrading the tier cannot cure a wifi or congestion problem.
Choose plans by evening speed
When you genuinely need a new plan, ignore your one-off test and compare retailers on the Typical Evening Speed, leaning on the ACCC Measuring Broadband Australia results. Most plans are month to month, so a congesting retailer is easy to leave.
Sizing a new plan? Read how NBN speed tiers really work, then check whether home wireless broadband is an option at your address.
Current figures, last updated 2026-06-16
Indicative Australian internet figures for 2026. Sources: NBN Co wholesale pricing, the Australian Competition and Consumer Commission (ACCC) Measuring Broadband Australia programme, and Canstar Blue 2026. Real speeds vary by your wifi, device, connection technology and the time of day you test.
The bottom line
Why the speed test is a clue, not a conclusion
An internet speed test is a brilliant diagnostic and a terrible verdict. It reads one path to one server at one instant, so it tells you whether your gear is working right now, but only if you run it wired and off-peak to strip out your wifi and the peak crowd. It cannot tell you which plan to buy, because the speed a plan delivers in daily life is the evening congestion a single test never sees. So treat a low result as a clue to investigate, not a reason to spend: diagnose your wifi, your device and the time of day first, fix what is actually broken, and when you do need a new plan, let the Typical Evening Speed and the ACCC results decide. In 2026, with most plans on no lock-in, that approach saves you from the most common and most expensive internet mistake of all: upgrading the line to fix a problem that was never the line.
Common questions
A Selectra expert answers your speed test questions
In most homes the answer is not the line, it is everything between the test server and your device. A speed test measures the whole path: your wifi, your device, your home network and the route to one test server, all at one moment. If you run it over wifi, the result reflects your wifi, not your NBN plan, and that gap can be large in a brick house or a long way from the router. The other big culprit is timing: run it at 8pm and you are also measuring peak-hour congestion. To judge the line itself, plug in by ethernet and test off-peak; only then are you close to measuring the plan rather than your loungeroom.
Not directly. A single test measures the slowest link in a chain, and that link is rarely the line. Your wifi, an old device, a busy home network, a VPN or the distance to the test server can all cap the result below what your plan can deliver. Even a clean wired test only tells you the speed to one server at one instant, not the speed you will get every evening. The honest number for what a plan delivers in daily use is the Typical Evening Speed the retailer publishes, because it already accounts for congestion across the busy hours.
Three rules make a speed test meaningful. First, plug your device into the router with an ethernet cable, so you are testing the line and not your wifi. Second, close other apps and pause downloads, and ideally test on one device with the rest of the household offline. Third, run it twice: once off-peak in the middle of the day and once during the busy 7pm to 11pm window, then compare. A healthy wired off-peak result that collapses at 8pm points squarely at peak-hour congestion, which is a retailer issue, not a line fault.
Latency, or ping, is the time in milliseconds for a signal to make the round trip to the server, and lower is better. Under 20 ms is excellent, under 50 ms is fine for almost everything including gaming, and over 100 ms starts to feel laggy in real-time apps. Ping matters far more than raw download speed for gaming and video calls, which is why a 1000 Mbps plan with poor latency can feel worse for a gamer than a modest plan with a clean, low ping. Wifi, VPNs and distance to the server all inflate ping, so test wired before you blame the connection.
Usually no, and this is the most expensive mistake the speed test causes. A slow reading is most often wifi, a device or peak-hour congestion, none of which a bigger tier will fix. Diagnose first: run a wired, off-peak test. If that result is healthy and only the wifi or the evening is slow, upgrading the tier is money spent on the wrong problem. The fixes that actually work are better wifi placement, a wired connection, or switching to a retailer with a higher Typical Evening Speed on the tier you already have.
Because the whole neighbourhood is online at once. Retailers buy a shared pool of bandwidth from NBN Co and decide how much to buy per customer. In the quiet daytime there is plenty to go around; in the 7pm to 11pm peak, demand surges and a retailer that bought too little lets everyone slow down. That congestion is a business decision, not a fault in the fibre, and it shows up in one published number: the Typical Evening Speed. If your wired off-peak test is fine but evenings crawl, the cure is a different retailer, not a different tier.
