Hash Lab

Cryptographic

BLAKE3

A successor to BLAKE2, released in 2020 by O’Connor, Aumasson, Neves, and Wilcox-O’Hearn. Built around a Merkle-tree of small BLAKE2-style compressions; the result is the fastest secure hash function on modern multi-core CPUs and a clean basis for streaming, parallel, and extendable-output use cases.

At a glance

OutputAny length (default 256 bits)
Block size1024 bits
ConstructionBinary Merkle tree of BLAKE2-style compressions
Rounds7
StandardNone (community-driven, reference impl by the designers)
Collision security2128 at 256-bit output
Length extensionNo
StatusModern, recommended

Three modes in one design

  1. Hashing , the default, plain content addressing.
  2. Keyed hashing , pass a 32-byte key to get a one-shot MAC, no HMAC needed.
  3. Key derivation , derive_key(context, input) uses an ASCII context string to domain-separate keys derived from the same input.

All three modes share the same compression function, distinguished by flag bits inside each node.

Why it is fast

Built-in XOF

BLAKE3 produces unbounded output: after finishing the absorption tree, you can call finalize(N) for any N. The streaming property and the keyed-derivation mode make BLAKE3 a natural fit for KDFs (a replacement for HKDF in some new designs).

Where it is used

Security

BLAKE3’s round count is aggressive, justified by the tree structure and a margin smaller than SHA-3’s. No collision or preimage attack better than generic is known on full BLAKE3 at the time of writing. For ultra-conservative use, BLAKE2b or SHA-3 retain larger margins.

Try it

The multi-algorithm hasher includes BLAKE3-256. Run the comparatorto see BLAKE3’s throughput against SHA-256 and SHA-3-256.

References

Visualize

BLAKE3-256 on your input

11 bytes · 0-bit digest

Hex digest

Bit grid (0 bits, teal = 1, slate = 0)

Byte pixel art (0 bytes, hue = byte value mod 360°)

Avalanche , flipping the lowest bit of the first input byte changed 0 of 0 output bits

Test Your Knowledge

Quick MCQ check on this algorithm

Start Quiz →

AI Summary

Summarize this page in your favorite LLM