IPv6 Calculator

Expand an IPv6 address to its full 32-digit form, compress it to the canonical short form, and work out what a prefix contains. Paste an address with or without a prefix, with a zone index, in brackets with a port, or with an embedded IPv4 tail — all of them are accepted and normalised.

Compression follows RFC 5952 exactly, which matters more than it sounds. The rules are: lowercase hex, no leading zeros inside a group, :: replaces the longest run of zero groups, the first such run wins a tie, and a run of only one group is never replaced. Implementations that get these wrong produce addresses that work perfectly but do not compare equal as strings — which silently breaks allow-lists, log correlation and deduplication in ways that are very hard to trace.

Prefix arithmetic is done with arbitrary-precision integers, so the address counts are exact. A /64 holds 18,446,744,073,709,551,616 addresses; a floating-point calculator rounds that and gives you something ending in a run of zeros. The more useful planning number is also shown: how many /64 subnets a prefix contains, since /64 is the standard subnet size.

There is also an EUI-64 builder, which turns a MAC address into the interface identifier that stateless autoconfiguration derives from it — including the universal/local bit inversion that everyone forgets — and the reverse ip6.arpa name, which is thirty-two reversed nibbles and thoroughly unpleasant to write out by hand.