English ▾
Free, Fast & Secure Online Tools
7z is the archive format created for the 7-Zip program in 1999. It does the same two jobs as ZIP — bundling files and compressing them — and it does the compressing considerably better, for reasons worth understanding because they come with trade-offs.
Better compression. 7z uses LZMA2 rather than ZIP’s much older DEFLATE. On compressible content — text, code, logs, databases — the difference is typically 30% to 50% smaller than the same files as a ZIP.
Solid archives. This is the interesting one, covered below.
Real encryption, including filenames. 7z encrypts with AES-256 and offers to encrypt the file list as well. That second part matters: a password-protected ZIP usually leaves the names of everything inside readable without the password, so anyone holding the archive can see that it contains salary-review-2026.xlsx. 7z can hide that too.
ZIP compresses each file independently. A thousand files means a thousand separate compression jobs, and nothing learned from one is used on the next.
7z can instead treat the whole set as one continuous stream and compress it in one pass. Repetition that occurs between files is then found and removed, not just repetition inside each file.
The effect is dramatic on sets of similar files. A folder of a thousand invoices sharing the same template, a project full of source code, a year of daily logs — the files are nearly identical to each other, and a solid archive collapses that shared material once instead of a thousand times. It is common to see a solid 7z at a fraction of the equivalent ZIP on content like that.
On a folder of photographs or videos it changes nothing, because those files are already compressed and share no byte-level repetition. Which brings us to the limitation 7z inherits unchanged.
Better algorithm or not, already-compressed data does not compress again. JPEG, PNG, MP4, MP3, PDF and Office documents are all compressed at the point they are created, and 7z finds no more repetition in them than ZIP does.
So the gain from 7z is real on text-like data and essentially zero on media. If you are archiving photos and the goal is a smaller file, the format is not the lever — the images are.
The compression that comes from treating everything as one stream has two consequences, and both are genuine.
Extracting one file is slow. To reach a file in the middle of a solid block, the software must decompress everything before it. Pulling one document out of a large solid archive can take as long as extracting the whole thing.
Damage spreads. In a ZIP, a corrupted file is a corrupted file; its neighbours are unaffected because each was compressed separately. In a solid block, corruption partway through can make everything after that point unrecoverable.
That second point is the one to weigh for backups. If an archive is going to sit on a drive for years, or travel over an unreliable connection, the resilience of independent compression may be worth more than the saved megabytes. 7-Zip lets you limit the solid block size as a middle path — the set is compressed in chunks, so damage is contained to one chunk while most of the benefit survives.
This is the deciding factor more often than compression ratio.
Windows and macOS both open ZIP files with no software installed. Neither opens 7z out of the box. Sending a 7z to a colleague, a client or a support desk means sending them a file they may not be able to use, and the reply is a request for a ZIP.
So the division is straightforward:
LZMA2 is slower than DEFLATE, markedly so at the higher settings, and it uses substantially more memory while compressing. Decompression is comparatively quick.
That asymmetry suits archiving: you pay once when creating the archive and the cost is small every time you read it. It suits routine, repeated compression much less well.