The public README is the clearest concise source for what Plonky3 actually is: not one fixed proving system, but a toolkit of primitives for implementing polynomial IOPs, mainly for STARK-based zkVMs and potentially other PIOPs.
The README status matrix is analytically useful because it exposes the real modularity surface: field choices (Mersenne31, BabyBear, KoalaBear, Goldilocks), generalized Merkle commitments, FRI-based polynomial commitments, univariate STARKs, multiple FFT implementations, multiple code families, and several hash functions. That is a more useful decomposition than filing Plonky3 as a generic fast prover.
The README’s CPU-features section matters because it shows performance depends not only on proof-system design but also on explicit SIMD / target-feature choices such as BMI, AVX2, AVX-512, and NEON. That makes Plonky3 a useful comparison point for how prover performance gets governed by implementation curation and hardware assumptions.
Polygon’s launch post adds the strategic framing missing from the repo alone: Plonky3 is positioned as a less opinionated successor to Plonky2, intended to let developers mix and match proving components depending on whether speed, proof size, or some other tradeoff matters most.
The same launch post is useful because it names downstream adoption by Valida and SP1, which makes Plonky3 more than an internal library. It is already functioning as a lower-layer proving substrate beneath other branded zkVMs.
No standalone Plonky3 whitepaper surfaced in this pass. The repo README and Polygon launch post are the strongest current primary-source packet for the toolkit itself.
Why this entry belongs in the corpus
Plonky3 gives the corpus a lower-layer proving-toolkit baseline beneath zkVMs, prover networks, and application-facing proof brands already cataloged.
Its durable mechanism is modular proof-stack composition: field choice, hash choice, PCS choice, FFT choice, and hardware-specific optimization are first-class governance surfaces rather than hidden implementation details.
The reusable insight is that practical proving power can concentrate in the maintainers of the shared library/toolkit layer long before it shows up as a chain, zkVM brand, or prover marketplace.