# zkVM

**Disclaimer: This page can be outdated.**

Based on the research done by Porter Adams, core contributor to Matter Labs/ZKsync. ([See video here](https://www.youtube.com/watch?v=fW6O40Pq_bM))

<table data-full-width="true"><thead><tr><th width="155.80078125">zkVM Name</th><th width="203.875">Succinct SP1 Hypercube</th><th width="231.1328125">RISCZero</th><th width="248.14453125">Brevis Pico</th><th width="283.63671875">Lita’s Valida</th><th width="284.34375">a16z’s Jolt</th><th width="289.65625">Nexus zkVM3.0</th><th width="290.61328125">ZKSync’s Boojum2.0</th></tr></thead><tbody><tr><td><strong>ZK tech name(s)</strong></td><td>Plonky3</td><td></td><td>Plonky3 + more</td><td>Plonky3</td><td>Arkworks + Binius</td><td>Stwo</td><td></td></tr><tr><td><strong>Arithmetization</strong></td><td>AIR</td><td>AIR</td><td>AIR</td><td>AIR</td><td>R1CS</td><td>AIR</td><td>AIR</td></tr><tr><td><strong>Proofs</strong></td><td>STARK</td><td>STARK</td><td>STARK / CircleStark</td><td>STARK</td><td>Sumcheck</td><td>CircleSTARK</td><td>STARK</td></tr><tr><td><strong>Commitments</strong></td><td>FRI</td><td>FRI</td><td>FRI</td><td>FRI</td><td>Zeromorph</td><td></td><td>FRI</td></tr><tr><td><strong>Prime Field</strong></td><td>Baby-Bear</td><td>Baby-Bear</td><td>BabyBear / Mersenne31</td><td>Mersenne31</td><td></td><td>Mersenne31 </td><td>Mersenne31</td></tr><tr><td><strong>Memory Argument</strong></td><td><a href="https://github.com/succinctlabs/sp1/blob/5c8a50e08b48d22b88471f39f9cc45947ca3bf5c/book/static/SP1_Turbo_Memory_Argument.pdf">Multi-set Hashing</a></td><td></td><td></td><td></td><td><a href="https://eprint.iacr.org/2025/105">Twist and Shout</a></td><td></td><td>Based on <a href="https://eprint.iacr.org/2023/1115">Two Shuffled Make a RAM</a></td></tr><tr><td><strong>Instruction Set</strong></td><td>RISC-V</td><td>RISC-V</td><td>RISC-V</td><td>Lita ISA</td><td>Risc-V</td><td>Risc-V</td><td>RISCV-32 I+M</td></tr><tr><td><strong>Other thoughts</strong></td><td><p>Extremely popular<br></p><ul><li>Easy to get started</li><li>Good docs &#x26; examples</li><li>product focused</li><li>developer friendly</li></ul></td><td><p>One of the first general-purpose zkVMs<br></p><ul><li>robust codebase</li><li>zirgen language</li><li>organized/ well-documented</li><li>devs with long-term vision</li></ul></td><td><p>Focused on modularity<br></p><ul><li>supports multiple proving backends</li><li>enshrined app level co-processors</li><li>focused on dev ex</li></ul></td><td><p>The truly fastest<br></p><ul><li>custom instruction set architecture (ISA)</li><li>modified from RISC-V to be more zk-friendly</li><li>custom compiler (30x faster)</li></ul></td><td><p>Unique cryptography choices<br></p><ul><li>Sumcheck 👀</li><li>Just One Lookup Table (JOLT)</li><li>smallest + simplest codebase</li><li>most likely to get fully formally verified (sumcheck easier to verifier than, say, FRI)</li></ul></td><td><p></p><p>Goal of processing 1 trillion Hz/s<br></p><ul><li>Switched from folding schemes and elliptic curve groups (Nova/Hypernova/Jolt &#x26; Ristretto 255 for Nexus zkVM2.0) to Stwo</li><li>uses <a href="https://specification.nexus.xyz/">"Harvard architecture in<br>which the program being executed resides in a read-only memory space separate from the data."</a></li></ul></td><td><p></p><p>Make devs happy<br></p><ul><li>separated proof system from the execution layer</li><li>Full EVM equivalent</li><li>should come out summer 2025</li><li>fast and cheap tx for users</li><li>4th full re-write of zk proof sytem</li></ul></td></tr></tbody></table>

[ETH proofs! - a site to determine who is proving ETH blocks for the cheapest cost and the fastest](https://ethproofs.org/)

Other projects not covered here include:

* Linea’s zkEVM
* Polygon’s Miden
* Starknet’s Cairo
* Scroll’s Ceno
* Aztec (needs client-side proving for privacy)
* Projects that modularize type of VM to use:
  * powdr
  * OpenVM