Both cables carry Gigabit Ethernet to 100 meters. That is the fact most comparisons step straight past, and it is why the honest answer to "which one should I use" is not simply "Cat6".
Figure 1: Comparison of internal conductor geometry between cable generations.The real difference is headroom. Cat5 is specified to 100 MHz of bandwidth. Cat6 is specified to 250 MHz. At 1G that gap barely shows, because both have margin to spare. At 2.5G, 5G and 10G the encoding gets denser, the cable has to keep four pairs quieter than Cat5 was ever asked to, and the gap becomes the whole story.
Quick answer
Question | Cat5 | Cat6 |
|---|
Bandwidth | 100 MHz | 250 MHz |
|---|
Gigabit to 100 m | Yes | Yes |
|---|
2.5GBASE-T | Not specified | 100 m |
|---|
5GBASE-T | Not specified | 100 m |
|---|
10GBASE-T | Not specified | 55 m |
|---|
Typical conductor | 24 AWG | 23 AWG |
|---|
Separator between pairs | None | Spline on most constructions |
|---|
Listed in the current TIA-568.2-D | No | Yes |
|---|
Why 150 MHz of extra bandwidth changes anything
Bandwidth in a copper cable is not a speed. It is the highest frequency the cable can carry while still meeting its crosstalk and return-loss limits — the point past which what arrives at the far end is no longer clean enough for the receiver to decode reliably.
Cat5 was written around 100 MHz and the limits that went with it. When Gigabit Ethernet arrived, it turned out that four-pair transmission at that rate loads all four pairs in both directions at once, which is a harder job than the signaling Cat5 had originally been designed around. Existing Cat5 cable often worked, but the margin was thin and depended heavily on how well the installer had terminated it.
Cat5e was the response: the same 100 MHz, with tightened near-end crosstalk, far-end crosstalk and return loss. That is what the "e" bought — not more bandwidth, but cleaner limits at the same bandwidth.
Cat6 went further and raised the ceiling. 250 MHz, plus a physical change: most Cat6 constructions run a spline down the center of the cable to hold the four pairs apart and keep coupling between them down. The conductor is usually 23 AWG rather than 24 AWG. Those two changes are what make 10GBASE-T possible to 55 meters, and what make 2.5G and 5G work over the full 100-meter channel.
The 100-meter channel, and where Cat5 runs out of room first
The 100-meter figure is a channel, not a reel length. It is made up of up to 90 meters of permanent link — the cable from the patch panel to the outlet, terminated once at each end — plus up to 10 meters of patch cordage split between the panel and the desk.
That split matters when you are judging whether an old installation has room to grow. A 92-meter Cat5 run that tests clean at 1G has almost no vertical margin left, and the first thing to suffer when somebody adds a longer desk lead is the crosstalk budget rather than the attenuation. If you are documenting an existing building, measure the permanent link separately from the cords.
The naming problem you will run into
A large amount of cable sold and installed as "Cat5" is actually Cat5e, because the jacket legend reads CAT5E and whoever wrote the quote shortened the name. If you are trying to work out what is inside your walls, read the printed legend rather than the purchase order.
The other thing worth knowing: Cat5 is no longer one of the categories the current TIA-568.2-D lists among its recognized grades. Cat5e is the floor. That does not stop Cat5 from working, but a new installation specified as "Cat5" is specifying a grade the standard no longer recognizes, and that is worth settling before the order goes out.
Does Cat5 still work? Usually, yes
For 1G to a desk, a phone, a printer or a camera, a healthy Cat5 link that passes its tests will run Gigabit. Failures are less about the category than about condition and termination:
- Crosstalk margin. Cat5’s limits were tight to begin with. A link that passed at handover can drift out of specification after a re-termination, a patch panel change or a poorly seated plug.
- Bundling. Cables run tight in a long bundle heat each other and couple into each other. A single run that tests clean in isolation can fail inside a 40-cable bundle, and alien crosstalk — coupling from neighboring cables rather than between pairs inside one cable — is not something a single-link tester sees.
- PoE. Both categories behave similarly at 15.4 W. Above that, heat inside the bundle becomes the limiter, and the older cable is usually the one carrying the thinner conductor.
Where each one is still a reasonable choice
Job | Reasonable choice | Why |
|---|
New horizontal cabling, 1G now, 10G later | Cat6 | 250 MHz keeps the 10G option open to 55 m without a second re-cable |
|---|
One office drop, 1G, existing Cat5 already in place | Cat5, if it tests clean | Replacing a working link usually costs more than the link is worth |
|---|
Access points on 2.5G or 5G | Cat6 | Both rates run the full 100 m on Cat6, and this is where NBASE-T ports actually land |
|---|
CCTV over PoE | Cat5e or Cat6 | Both carry 1G and the PoE load; the deciding factors are run length and bundle size |
|---|
Three field scenarios
A 2004 office with Cat5 to every desk. The links test and the tenant is on 1G. Re-cabling is not justified by the network — but it may be justified by the ceiling. If a fit-out is already open, pulling Cat6 now costs a fraction of pulling it later.
A home with 1G fiber and a handful of drops. Cat5e handles this with room to spare. The upgrade worth doing is not the cable but the terminations; a badly seated keystone costs more throughput than the category does.
A warehouse adding 2.5G access points at the far end. This is where Cat5 loses. 2.5GBASE-T is not specified over Cat5, and the runs sit at the length where margin matters most.
What the price difference is really buying
Cat6 costs more for three reasons: a larger conductor, tighter pair geometry that takes more process control at the extruder, and — on shielded constructions — a foil or braid layer plus a drain wire. None of that is marketing. All of it is what buys the extra 150 MHz.
Where it stops being worth it is a short, healthy existing run. Where it is clearly worth it is any new installation where the conduit or ceiling void will not be opened again for a decade.
FAQ
Is Cat6 faster than Cat5 at Gigabit?
No. Both carry 1G to 100 meters. Cat6 carries it with more margin and can go beyond 1G; Cat5 cannot.
Can I mix Cat5 and Cat6 in the same network?
Yes. A channel takes on the performance of its weakest part, but a mixed network runs fine as long as each individual channel meets the speed you need.
Do I need Cat6 if I only have a 1G switch today?
If the cable is being installed now, yes. Labor costs more than the cable, and a switch upgrade is a five-minute job compared with a re-cable.
Does the "e" matter?
Yes, but as margin at the same bandwidth. Cat5 and Cat5e are both 100 MHz; Cat5e is the version whose limits were written for four-pair Gigabit.
How do I tell which one I have?
Read the jacket legend. The category, gauge and temperature rating are printed on the cable at regular intervals.
Will Cat6 make my internet faster?
No. It raises the ceiling on what your local network can carry. The speed you buy from an ISP arrives on the same fiber or coax either way.
Related reading on this site
