Jimmy Liao, Neo core developer and founder of R3E Network, published two experimental repositories on May 4 exploring what Neo’s next generation could look like. The larger of the two, neo-n4, prototypes a multi-L2 elastic network architecture built on Neo 4 core. The second, neo-lang, is an early-stage domain-specific language for Neo N3 smart contracts.
Both projects are independent community research efforts. The neo-n4 repository prominently states that it is “NOT the official Neo 4 release,” describing itself as “one community’s prototype, not a spec.”
neo-n4: elastic network architecture
neo-n4 is envisioned as a three-tier design where Neo N3 or Neo 4 serves as the L1 settlement layer, an optional gateway aggregates proofs from multiple L2 chains, and individual L2 chains run Neo 4 core as their execution kernel. The architecture borrows the shared-bridge and proof-aggregation pattern pioneered by ZKsync’s Elastic Chain, rebuilt on Neo’s stack using dBFT 2.0 finality, NEP-17 assets, NeoVM, and NeoFS for data availability.
The scope of the prototype is substantial. At the time or writing, the repository contains 820 tests across 26 projects, 19 smart contracts split between 13 NeoHub L1 contracts and 6 L2 native contracts, 15 off-chain libraries, 8 node plugins, and 3 CLI tools. Contracts cover chain registration, shared bridging, settlement management, sequencer bonding, and an optimistic challenge window, among other functions.
Liao structured the work in phases. Phases zero through three, covering a sidechain proof of concept, the NeoHub shared bridge, batch settlement, and an optimistic challenge window, are marked as complete. Phase six is also complete, providing developer CLI tooling. Phases four and five, which target ZK validity proofs using a RISC-V prover and proof aggregation across multiple L2s remain in progress with scaffolding in place.
The project’s cross-chain messaging system, Neo Connect, outlines L1-to-L2, L2-to-L1, and L2-to-L2 message passing through batched Merkle proofs. A tiered data availability model offers three levels: L1 settlement for high-security use cases such as DeFi, NeoFS for lower-cost applications, and a data availability committee option for minimal-cost scenarios.
The repository includes a whitepaper, architecture documentation, and operator guides, though it has not undergone any disclosed security audit.
Relationship to official Neo 4 work
neo-n4 explicitly assumes Neo 4 core exists as a working base layer and builds an L2 architecture on top of it. The project’s fourth phase targets NeoVM 2 and RISC-V ZK validity proofs, aligning with the direction Neo co-founder Erik Zhang outlined in his draft Neo 4 roadmap in September 2025.
On April 15, Zhang announced that a NeoVM-compatible RISC-V VM solution had passed full MainNet state validation, confirming the design had moved beyond the conceptual stage. Liao contributed to that effort, sharing an architecture diagram of PolkaVM integration into the Neo Core C# node days earlier.
The neo-n4 announcement came 19 days after that milestone. Where Zhang’s work represents canonical Neo 4 protocol development, neo-n4 explores what a multi-L2 scaling layer could look like if built on top of it. Zhang’s draft roadmap did not explicitly propose a multi-L2 architecture.
neo-lang: a Neo-native contract language
The second repository, neo-lang, presents a contract-oriented language targeting Neo N3 with a `.neo` file extension. It features 10 built-in types, struct declarations, event handling, a package system, and access to Neo native contracts including Oracle and Notary.
Liao’s announcement claims the language saves 30% on opcodes. However, while the repository contains a comprehensive language reference, compiler binary, test suites, example contracts, and benchmarks are not yet available.
The contrast with R3E’s more mature neo-solidity compiler, which shipped v0.15.0 on March 20 with over 700 tests, example DeFi contracts, and Hardhat integration at 95% completion, is notable. Both projects target Neo N3 bytecode but serve different audiences. neo-solidity ports an established language for EVM-familiar developers, while neo-lang is aimed at being Neo-native from the ground up.
R3E’s prolific output continues
The two repositories are the latest in a rapid stream of developer tooling from R3E in 2026. Since February, the team has shipped the neo-solidity compiler, a JavaScript decompiler SDK, a TEE-powered oracle system deployed to MainNet, and SDK releases for JavaScript, Rust, and Swift.
neo-n4 represents the most ambitious of these efforts – a prototype of network-level infrastructure rather than individual developer tools. Whether any of its components find their way into official Neo development remains an open question. Liao’s dual role as both R3E founder and Neo core developer positions the work as informed exploration, but the repository’s own disclaimers make clear it is exactly that – an exploration, not a roadmap.
The neo-n4 and neo-lang repositories can be found at the links below:
https://github.com/r3e-network/neo-n4
https://github.com/r3e-network/neo-lang