Start with the function
A single list of “quantum-resistant algorithms” can hide the distinction that matters most to an application. A KEM establishes shared key material. A signature lets others verify that a holder of a private key signed a message. Upgrading a secure connection and upgrading blockchain authorization therefore require different tools.
The table describes functions and publication stages at the stated review date. It is a navigation aid rather than a recommendation for a production parameter choice.
| Algorithm | Function | Family | Publication stage |
|---|---|---|---|
| ML-KEM | Key establishment | Module lattices | FIPS 203 · Final |
| ML-DSA | Digital signature | Module lattices | FIPS 204 · Final |
| SLH-DSA | Stateless digital signature | Hash-based | FIPS 205 · Final |
| Falcon / future FN-DSA | Digital signature | NTRU lattices | Selected; FIPS 206 in development |
| XMSS | Stateful digital signature | Hash-based | RFC 8391; specified sets in SP 800-208 |
| HQC | Key establishment | Code-based | Selected; standard in development |
Compare operational assumptions
XMSS requires one-time signing state to be managed across signing, backup and recovery. A stateless signature scheme does not have that same consumed-index requirement. Both still require secure key storage and a correct surrounding transaction protocol.
For a system design, ask who signs, how many devices may sign concurrently and what happens after a crash. A signer that can only be recovered from an old backup has different requirements from a controlled signing appliance. Those scenarios deserve explicit tests; they cannot be inferred from a mathematical family label.
Sources:NIST SP 800-208
Measure like with like
For signatures, record the selected parameter set, public-key encoding, signature encoding and verification work. For KEMs, record encapsulation-key, ciphertext and shared-secret handling. Those are different data categories. A ciphertext is not an alternative kind of blockchain signature.
Our comparison method requires a common environment before publishing speed claims: library revision, processor, compiler, build options and measurement procedure. A chart built from unrelated vendor numbers can describe the sources, but it does not establish which construction is fastest for a particular wallet or node. This page intentionally makes no measured performance claim.
Sources:NIST SP 800-227
Add protocol and deployment evidence
A signature primitive does not decide which transaction bytes are signed or which network accepts a new format. A KEM does not, on its own, define the authentication policy for a connection. Our recommended evaluation records those protocol rules next to the implementation revision.
The evidence chain should identify a specification, a matching implementation, tests, an activation rule and the scope actually deployed. Source availability, a testnet demonstration and mainnet activation are useful but distinct observations. Preserve them as separate fields rather than collapsing them into a general safety score.
Sources:NIST SP 800-227
How to use this library
Open an algorithm page for its role and sources, then follow the relevant project profile. When a project says it uses Falcon or Dilithium, check its exact version and deployment scope before relating it to a standardized name. Publication stages can change, so use the review date as a cue to recheck the owning source.
CoinTalk has not audited the compared implementations or tested their resource use. The library provides questions and references that make a review more precise, with the technical decision left to a protocol-specific assessment.
Sources:NIST PQC standardization process