--- title: "Sui vs Aptos: Move-based Chains, Object Model vs Account Model" description: "Comparing Sui and Aptos on data models, speed, DeFi, and developer tools" author: "Sui Foundation" datePublished: 2026-10-27 dateModified: 2026-10-27 category: "Move" readTime: "8 min" canonical: https://www.sui.io/blog/sui-vs-aptos-move-object-account-model --- # Sui vs Aptos: Move-based Chains, Object Model vs Account Model ## Main Takeaways - Sui and Aptos both came out of the Diem project at Meta and both use Move, so a developer who learns the language can read code on either chain. The divergence is the data model, and the rest of the comparison follows from it. - Sui is object-centric, so the chain knows before execution exactly which objects a transaction touches, and simple operations on owned objects can settle on a fast path without waiting for full consensus. Aptos is account-centric and recovers parallelism at runtime with Block-STM, which runs transactions speculatively and re-executes the ones that conflict. - Both chains are shipping. Aptos has Block-STM V2, the Raptr consensus protocol, and sub-50-millisecond block times. Sui has Mysticeti v2, which integrated validation into consensus and cut measured latency by 25 to 35 percent depending on region. - DeFi liquidity is the clearest gap. Sui holds about $473 million in DeFi TVL against roughly $56 million on Aptos (DefiLlama, September 2026). Both are well below their spring peaks, so check current figures before deciding. Both Sui and Aptos came out of the Diem project at Meta, both use the Move programming language, and both were built by engineers who had spent years trying to build global payment infrastructure inside a large company. That shared origin makes the comparison unusually clean, because the two chains agree on a great deal and disagree on a few specific decisions that turn out to matter. ## The shared inheritance What the two chains have in common is substantial. Both use Move, a language designed from the start to treat assets as first-class, safety-protected resources. Both declined to inherit Solidity's exploit surface. Both pursue parallel execution rather than the single-threaded processing of older chains. A developer who learns Move can read code on either chain. The divergence is in the data model, and from that one decision a chain of consequences follows. Sui adopted an object-centric model. Aptos kept an account-centric model closer to Ethereum's structure. Everything else in this comparison traces back to that fork. ## Object model versus account model In Sui's object model, every asset is a distinct on-chain object with its own ID, its own type, and explicit ownership rules. Because each asset is a separate object, the chain knows before execution exactly which objects a given transaction will read and write. Transactions over different objects run in parallel with no coordination, and simple operations on owned objects can settle on a fast path without waiting for full consensus. Parallelism is a property of the data itself. Aptos uses an account model and achieves parallel execution through Block-STM, an optimistic concurrency approach. Block-STM runs transactions in parallel speculatively, detects conflicts where two transactions touched the same state, and re-executes the ones that clashed. It is an effective technique, and it lets Aptos parallelize without requiring the object structure. Block-STM V2, on the Aptos roadmap for 2026, is designed to make throughput scale linearly with validator core count rather than plateau. The structural cost stays the same: parallelism is recovered at runtime through conflict detection and re-execution, so contended state pays for its contention twice, once to execute and once to redo. Both designs work. Block-STM is flexible and keeps a familiar account model, which also makes cross-chain compatibility with EVM ecosystems more straightforward. The object model makes transaction independence explicit, which is what gives Sui its very high per-account throughput and its consensus-bypass fast path for simple transfers. For applications with complex, composable ownership, where many assets move independently, the object model expresses the logic more directly. ## Both chains are still shipping Aptos has not stood still. It reached sub-50-millisecond mainnet block times by December 2025, the fastest block production among major Layer 1s. Raptr, its new consensus protocol, targets sub-second finality under network stress. Block-STM V2 is in refinement, and X-Chain Accounts are on the 2026 roadmap for interoperability with Solana, Ethereum, and other ecosystems. One distinction is worth keeping straight when reading those numbers. Block time measures how often a chain produces a block. Finality measures when a transaction becomes irreversible. Sui's 390 milliseconds is deterministic finality. Aptos's sub-50 milliseconds is block production, with Raptr aiming at sub-second finality. On the Sui side, Mysticeti v2 integrated transaction validation into consensus, so the commit rule applies to every transaction concurrently and transactions are finalized or rejected alongside block decisions on the same DAG. A new Transaction Driver replaced the Quorum Driver, so a client submits to one chosen validator and retrieves effects with quorum validation afterward. Measured latency fell about 35 percent for Asia-based nodes and about 25 percent for Europe-based nodes. ## Where the ecosystems stand Shared heritage has not produced identical traction. Sui's DeFi total value locked stands at about $473 million against roughly $56 million on Aptos (DefiLlama, September 2026). Both numbers are far below their spring highs, when Sui peaked near $2.6 billion and Aptos near $1 billion, so read them as a snapshot of a drawdown rather than a steady state. The ratio has held or widened in Sui's favor through it. Developer counts point the same way. In Q1 2026 Sui had roughly 954 monthly active developers against about 465 on Aptos, a little over two to one. Aptos carries breadth of its own: about 330 ecosystem projects across roughly 30 active DeFi protocols, a more distributed structure that some institutional risk teams prefer, plus a stablecoin market cap in the $1.5 to $1.8 billion range and about $1.2 billion in tokenized real-world assets. Liquidity depth compounds. Deeper DeFi liquidity attracts more applications, which attract more users and more liquidity. For a builder choosing where to deploy a financial application, the size and activity of the existing ecosystem is a practical consideration alongside the architecture. ## The stack beyond the base chain Both chains are more than their consensus and execution layers, and they have invested in different surrounding capabilities. Sui ships an integrated stack: Walrus for decentralized storage, Seal, which encrypts data and manages the access policies that decide who can decrypt it, Nautilus for trusted off-chain computation, DeepBook as a native on-chain liquidity layer with an order book, margin, and prediction primitives, gasless stablecoin transfers, zkLogin for one-step sign-in with an existing Google or Apple account, and confidential transfers, in public beta on devnet, which shield transaction amounts from public observers. Sui Dollar, issued by Bridge, a Stripe company, extends the stack toward payments, and Hashi, which reached testnet in July 2026, extends it toward institutional Bitcoin in DeFi. These pieces were designed to compose with one another. Aptos has its own growing ecosystem and tooling and continues to ship, with framework-level central limit order book infrastructure among its 2026 priorities. The distinction for a builder is the degree of first-party integration. Sui has concentrated on shipping a composable stack as part of the platform, so more of what an application needs is available natively. ## Sui and Aptos at a glance Dimension Aptos Sui Origin Diem project, Move language Diem project, Move language Data model Account model Object model Parallel execution Block-STM optimistic concurrency; V2 in refinement Object independence, known before execution Per-account throughput Contended state pays for conflict detection and re-execution Very high; independence is structural Simple transfers Standard execution path Can settle on a fast path without full consensus Consensus Raptr, targeting sub-second finality Mysticeti v2, validation integrated into consensus Block time and finality Sub-50ms block time (Dec 2025); Raptr targets sub-second finality ~390ms deterministic finality DeFi TVL (Sept 2026) ~$56 million ~$473 million Monthly active developers (Q1 2026) ~465 ~954 Integrated stack Growing ecosystem, framework-level CLOB in progress Walrus, Seal, Nautilus, DeepBook native Swipe or scroll horizontally to compare all columns. _TVL: DefiLlama, September 2026. Developer counts: Q1 2026. Aptos roadmap figures: Aptos Labs and Everstake, 2026._ ## Which one should you build on If your team prefers an account model close to Ethereum's mental model while still writing in Move, wants the cross-chain compatibility that model makes easier, or has specific reasons to favor Block-STM's approach to concurrency, Aptos is a credible and well-engineered choice. The two chains share enough that the decision is rarely about which is broadly better. Sui is the stronger choice when you want transaction independence to be explicit rather than recovered at runtime, when per-account throughput matters for agent-driven or high-frequency workloads, when you want the fast path for simple transfers, or when the deeper DeFi liquidity and the integrated first-party stack shorten your path to a working product. For applications with rich, composable ownership, the object model is the reason builders are choosing Sui. ## FAQ **Are Sui and Aptos the same underneath?** They share an origin in the Diem project and both use the Move language, so they are close relatives. The core difference is the data model: Sui is object-centric and Aptos is account-centric, which leads to different approaches to parallel execution and different performance characteristics. **What is the difference between Block-STM and Sui's object model?** Block-STM runs transactions in parallel optimistically and re-executes any that conflict, recovering parallelism at runtime over an account model. Sui's object model knows before execution which objects a transaction touches, so non-overlapping transactions run in parallel without conflict detection or re-execution. **Which has more DeFi activity, Sui or Aptos?** Sui's DeFi total value locked is about $473 million against roughly $56 million on Aptos as of September 2026, and Sui carried roughly twice Aptos's monthly active developer count in Q1 2026. Both TVL figures are well down from spring peaks, so verify current numbers before deciding. **Is Aptos faster than Sui?** The two figures usually quoted measure different things. Aptos reached sub-50-millisecond block times in December 2025, which is how often blocks are produced. Sui settles in roughly 390 milliseconds, which is when a transaction becomes irreversible. Aptos's Raptr consensus targets sub-second finality. **Do both chains use Move?** Yes. Both Sui and Aptos use Move, the asset-safety-focused language from the Diem project. The implementations differ because Sui's Move works with an object model while Aptos's works with an account model. ## Sources and further reading - Sui object model and Move at [docs.sui.io](https://docs.sui.io) - [Mysticeti v2: Faster and Lighter Sui Transaction Processing](https://www.sui.io/blog/mysticeti-v2-sui-consensus) - [Aptos in 2026: Latest News, Roadmap, and DeFi Updates](https://everstake.com/resources/blog/aptos-news-2026) for Block-STM V2, Raptr, block times, and ecosystem figures - [Aptos vs Sui: Competitive Landscape of the Move Duo](https://www.gate.com/blog/aptos-vs-sui-move-ecosystem-institutional-defi-2026) for Q1 2026 developer counts and peak TVL - [DefiLlama chain TVL](https://defillama.com/chains) for current DeFi totals - DeepBook, Sui Dollar, and Hashi at [sui.io/blog](https://www.sui.io/blog)