Sui vs Aptos: Move-Based Layer 1s Compared

Introduction

Sui and Aptos share a parent, Meta's abandoned Diem project, yet the two chains measured on 5 November 2026 look very different. Sui holds $462.7M in decentralized finance total value locked against $40.8M on Aptos, while Aptos holds $779M in stablecoins against $521M on Sui (DefiLlama). SUI sits 82.7% below its all-time high and APT 96.8% below, and Sui has halted five times since November 2024. This head-to-head compares the two Move dialects, consensus designs, measured throughput, supply schedules, ecosystem data and outage records. It separates live upgrades from roadmap items and vendor claims from measurements, so a reader can judge which chain fits a given job and which numbers will need rechecking soon, since every figure carries its date.

How Do Sui Move and Aptos Move Differ at the Data-Model Level?

Sui and Aptos, two monolithic layer 1 blockchains ↗, both descend from Diem, Meta's abandoned payments blockchain, and both run the Move language. They diverge at the data model: Sui rebuilt storage around objects with unique IDs, while Aptos kept Move's account-and-resource storage.

From Diem to two Move dialects

Move was written for Diem, which Meta launched as Libra, and its account-based data model came with it (Sui Foundation, 2026). Mysten Labs forked Move into Sui Move with an object-based model. Aptos Move builds on the original Move created by the Diem and Novi teams (Messari, 2025) and keeps global-storage operators such as move_to and borrow_global (Sui docs, 2026). The dialects now differ enough that Mysten Labs' own developer FAQ tells builders not to apply Aptos patterns, such as signer-based ownership, on Sui (Mysten Labs, 2026). Libraries and audits therefore do not transfer between the chains. Single-chain guides cover how Sui works and what the SUI token does ↗ and how the Aptos blockchain and its Move language work ↗ on their own.

Objects versus accounts and resources

Sui stores every asset as an object with a unique ID, an owner and a version number (Sui docs, 2026). Each transaction names all its inputs up front, so the network can run transactions with non-overlapping inputs in parallel (Sui docs, 2026). Aptos stores resources under accounts, and a transaction reaches them through global storage operations. Sui also splits ownership types: an address-owned object takes the fast path, while a shared object such as a liquidity pool needs consensus ordering and adds latency (Sui docs, 2026). Builders therefore design around contention on Sui and around account state on Aptos.

Data current as of November 2026.

Parallelism follows from these models, but each chain pairs it with a consensus design that has changed since launch.

Which Consensus and Execution Upgrades Are Live on Sui and Aptos Today?

Sui's Mysticeti v2, Aptos's Baby Raptr and Velociraptr, and Aptos's Block-STM v2 are live (Mysten Labs, Aptos Network and aptos-core, 2025-2026). Full Raptr and Archon have no mainnet activation notice in Aptos Labs' posts or release notes (Aptos Labs, 2025). The table below separates shipped upgrades from research and roadmap items as of 5 November 2026.

Sui Mysticeti and the owned-object fast path

Mysticeti, a directed-acyclic-graph consensus protocol, reached Sui mainnet in July 2024, and Mysticeti v2 followed in November 2025 (Mysten Labs, 2025). Version 2 folds transaction validation into consensus and adds a Transaction Driver that sends each transaction to a single validator. Mysten Labs reported latency cuts of 35% on Asia-based full nodes, from about 1.00 to 0.65 seconds, and 25% in Europe, from 0.55 to 0.40 seconds, which are vendor measurements. Transactions touching only address-owned objects skip full consensus, while shared-object transactions wait for ordering (Sui docs, 2026). Mysten Labs names cutting the median message rounds needed to commit a transaction from four to three as the next phase (Mysten Labs, 2025). Sui mainnet reported protocol version 139 at checkpoint 341,906,079, at 12:00 UTC on 5 November 2026 (Sui GraphQL, 2026).

Aptos Quorum Store, Baby Raptr and Velociraptr

AptosBFT pairs the two-chain Jolteon protocol with Quorum Store, a layer where every validator broadcasts transaction batches in parallel, and blocks with fewer than 100 transactions bypass it (Aptos Network, 2025). Baby Raptr (Aptos Improvement Proposal AIP-106) went live on mainnet in June 2025 and cut the path from six network hops to four, which Aptos said improved validator finality latency by 20%, or 100 to 150 milliseconds (Aptos Network, 2025). Velociraptr (AIP-131) lets leaders propose a block before the parent's certificate arrives, and its rollout ran from September to October 2025 (Aptos Labs, 2025). Aptos Labs claimed block times under 50 milliseconds (Aptos Labs, 2025), and a 5 November 2026 sample measured about 37.5 milliseconds, from 26.7 blocks per second (Aptos fullnode API, 2026).

Execution engines and what is still roadmap

Aptos' Block-STM executes transactions optimistically in parallel and re-executes those that conflict (Aptos Labs, 2022). Release v1.49.1 of aptos-node, published 23 September 2026, lists "Enable Block-STM v2 by default" (aptos-core, 2026), and CryptoBriefing reported Aptos' confirmation on 15 October 2026. Two Aptos items remain unproven on mainnet. The Raptr paper reports 260,000 transactions per second in a 100-replica lab test, with 755 milliseconds latency at 250,000 transactions per second (Aptos Labs, 2025). Archon, a design targeting 10-millisecond blocks, appeared as an anticipated 2026 proposal (Aptos Labs, 2025). No mainnet activation notice for either had turned up by 5 November 2026.

Data current as of November 2026.

Upgrade notes describe latency and capacity, while chain data shows what the networks process.

How Fast Are Sui and Aptos When Throughput Is Measured, Not Claimed?

On 5 November 2026 Sui processed about 156 transactions per second and Aptos about 171, measured from each chain's public endpoints. Both sit orders of magnitude below the 100,000-plus transactions-per-second (TPS) figures in promotional material (CoinEx, 2024).

Measured throughput on 5 November 2026

Sui's GraphQL endpoint reports a cumulative transaction count and a timestamp for every checkpoint. The 12 hours from 00:00 to 12:00 UTC averaged 156 transactions per second at 4.12 checkpoints per second (Sui GraphQL, 2026). Aptos' REST API reports block timestamps and ledger versions, and the same 12-hour window averaged 171 transactions per second at 26.7 blocks per second (Aptos fullnode API, 2026). Three-hour sub-windows gave 150 to 160 on Sui and 162 to 187 on Aptos, so the averages are reasonably stable. Both counts include system transactions, so user activity is lower, and neither window is a stress test.

Finality and vendor claims

Sui says transactions now finalize in well under a second for most use cases (Sui, 2025), and the Mysticeti v2 post reports full-node confirmation latency of 0.40 to 0.65 seconds. Aptos reports a 100 to 150 millisecond finality gain from Baby Raptr, and its Raptr lab test reached 755 milliseconds at 250,000 transactions per second, which is not mainnet. Third-party research in 2024 repeated 100,000-plus TPS capability for both chains (CoinEx, 2024), several hundred times the measured figures. A search on 16 October 2026 found no independent end-to-end finality benchmark for either chain. Block cadence, about 37.5 milliseconds on Aptos and 243 milliseconds on Sui, measures how often a chain produces output, not how long a user waits for finality.

Data current as of November 2026.

Throughput says little about value, so the next comparison is how much of each token exists and when more arrives.

How Do SUI and APT Supply, Unlocks and Drawdowns Compare?

SUI has 41% of a fixed 10 billion supply in circulation and releases more every month through 2030. APT has 884 million in circulation, and its four-year investor unlock cycle ended on 12 October 2026.

Supply, circulation and unlock schedules

SUI's supply is fixed at 10 billion, with about 5% circulating at the May 2023 launch (Sui, 2026). Tokenomist showed 4.14 billion SUI, or 41.37%, unlocked on 6 November 2026, after monthly cliff releases of 13.7 million SUI on 1 August, 13.5 million on 1 September, 13.3 million on 1 October and a further release on 1 November to the Community Reserve; the next is listed for 1 December (Tokenomist, 2026). Aptos minted 1 billion APT at mainnet and has paid about 199 million more as staking rewards (Aptos Foundation, 2026; CoinPaprika, 2026). Validators held 7.07 billion SUI and 767.6 million APT in stake on 5 November 2026 (Sui GraphQL and Aptos fullnode API, 2026), figures that include delegated and locked tokens and so cannot be read against circulating supply. Aptos' investor and core-contributor cycle ended with the 12 October 2026 release, cutting scheduled monthly releases from about 11.3 million to 4.5 million APT, and the next listed unlock, on 11 November, goes to Community and Foundation (Tokenomist, 2026; CryptoNews, 2026).

Drawdowns from all-time highs

SUI sits 82.7% below its January 2025 high of $5.35, and APT sits 96.8% below its January 2023 high of $19.95 (CoinPaprika, 2026-11-05). Aptos Foundation's AIP-140 proposes a 2.1 billion hard cap, a staking-rate cut from 5.19% to 2.6% and a permanent 210 million APT stake. The AIP repository lists it as Accepted, but enforcement of the cap was not independently confirmed (Aptos Foundation, 2026). Supply policy has not reversed the drawdown, and none of these figures predicts a future price; each records past price against a peak.

Data current as of November 2026.

Supply schedules matter less than whether anyone uses the chain, so ecosystem data comes next.

How Do the Sui and Aptos Ecosystems Compare on TVL, Stablecoins and DEX Volume?

Sui holds $462.7M in decentralized finance (DeFi) total value locked (TVL) against $40.8M on Aptos, about 11 to 1, while Aptos holds more stablecoins, $779M versus $521M. Both chains sit far below earlier peaks (DefiLlama, 2026-11-05).

TVL and what drives it

DefiLlama ranks Sui 15th among chains by TVL and Aptos 38th. Sui peaked at $2.64B on 9 October 2025 and Aptos at $1.31B on 17 December 2024, so they have fallen 82% and 97% (DefiLlama, 2026). On Sui, the lending markets Suilend ($146.9M) and NAVI ($138.4M) lead, followed by SpringSui liquid staking at $62.0M, the lending market Current at $56.7M, Haedal at $34.0M and Cetus CLMM, the exchange hit by the 2025 exploit, at $25.3M. A year earlier Sui held $1.42B and Aptos $581M, so 12-month declines reach 67% and 93%. On Aptos, Decibel, an on-chain trading venue, leads with $29.0M, and no other non-exchange protocol exceeds $6M, with the tokenized-fund platform KAIO at $5.7M and the lending market Echelon at $5.1M. Centralized exchange reserves are excluded.

Stablecoins on each chain

Aptos holds $778.7M in stablecoins, led by USDT at $337.0M and USDC at $265.6M, and DefiLlama counts BlackRock's BUIDL fund at $157.3M of that total, about 20% of the Aptos balance. The balance fell from $1.05B 30 days earlier, a $274M or 26% drop. Sui holds $520.6M, up 8% from $480.7M a month earlier, with USDC at $336.1M and USDSUI at $74.2M (DefiLlama, 2026). Aptos therefore leads on stablecoin balances despite a TVL one-eleventh of Sui's.

DEX volume and fees

Sui's decentralized exchanges traded $36.9M on 4 November 2026 and $892.7M over the 30 days to that date, led by Cetus CLMM at $13.1M, DeepBook V3 at $9.2M and Bluefin Spot at $3.8M that day. Aptos traded $3.7M on 4 November and $168.4M over 30 days, led by ThalaSwap V3 at $3.4M and Hyperion at $0.25M (DefiLlama, 2026-11-05), so Sui's exchanges handle roughly five times the volume over 30 days. Sui's seven-day total was $172.1M against $35.8M on Aptos. Concentration is high on both: Cetus alone accounts for about 35% of Sui's daily volume, and ThalaSwap V3 for about 90% of Aptos'. Chain fees reached $154K in 24 hours on Sui against $41K on Aptos. Dividing fees by the measured transaction counts gives roughly $0.000011 per transaction on Sui and $0.0000027 on Aptos, a rough derived estimate.

Data current as of November 2026.

Activity figures show demand, and the incident record shows how the networks behave when they fail.

What Incidents Have Hit Sui and Aptos, and What Do They Show?

Sui has halted five times since November 2024 and froze $162M during the Cetus exploit. Aptos has one primary-sourced mainnet halt, in October 2023, and no later post-mortem had turned up by 6 November 2026.

Sui outages and the Cetus freeze

Sui's mainnet stopped on 21 November 2024 when a congestion-control assertion crashed every validator, and on 14 January 2026 when a consensus divergence halted checkpoints for about six hours (Sui, 2026). Three more halts followed on 28 and 29 May 2026 after release v1.72 introduced address balances; no funds were at risk and no transactions were reverted (Sui, 2026). On 22 May 2025 an attacker drained about $223M from Cetus through a flaw in an open-source library, and validators froze about $162M (crypto.news, 2025). A vote on 29 May passed with 90.9% of stake, and the funds moved to a four-of-six multisig held by Cetus, the Sui Foundation and OtterSec (Sui Foundation, 2025). Critics said the freeze showed limits to decentralization (crypto.news, 2025).

Aptos incident record and validator concentration

Aptos halted transactions on 18 October 2023 from about 4:15 PM to 9:18 PM PDT, after non-deterministic code produced differing gas results; no committed transactions were lost and no fork occurred (Aptos Foundation, 2023). The trigger was a performance change merged on 22 August 2023 that interacted with the FeeStatement event, which went live on 16 October. Aptos called it the first significant delay since launch. A search on 6 November 2026 found no later mainnet halt post-mortem, which is not proof that none occurred. One third of voting power can stall ↗ a Byzantine fault tolerant (BFT) network, and the top 14 of Aptos' 83 active validators exceed that line, against 18 of Sui's 125 (Aptos and Sui endpoints, 2026).

Reliability history sets the risk, and the last question is which chain suits which job.

Which Chain Fits Which Builder or User?

Sui fits builders who model assets as objects and ship consumer or gaming flows, while Aptos fits teams that want account-style Move and a trading-engine focus. Both tokens remain far below their highs, so intended use should drive the choice.

Builders: language, tooling and execution fit

Sui Move requires thinking in objects, ownership types and shared-object contention, and its code does not port to Aptos (Mysten Labs, 2026). Aptos Move keeps the account-and-resource patterns of original Move. Sui's 2025 stack added Walrus storage, Seal access control and Nautilus on mainnet (Sui, 2025), while Aptos Labs builds around Decibel, an on-chain trading engine, and an encrypted mempool that went live as an opt-in on 12 October 2026 (Aptos Network, 2026). Both chains ship upgrades quickly: Aptos governance executed framework releases v1.46.0 on 15 June, v1.47.1 on 12 July and v1.48.0 on 16 August 2026 (Aptos governance, 2026), and Sui reached release v1.72 by May 2026, which also shows the regression risk behind its outages. Builders should map an app's contention pattern first: many independent owners favor Sui, and heavy shared-state trading logic is what Block-STM v2 targets (CryptoBriefing, 2026).

Users and investors: what the data supports

On this evidence Sui has the deeper DeFi base, a longer outage record and a fixed supply with about 59% still to unlock. Aptos has the larger stablecoin balance, one primary-sourced halt since 2023 and an investor unlock cycle that ended in October 2026. Transaction costs are a minor factor, since the derived average fee is a fraction of a cent on both chains. Neither dataset supports a price view, since SUI sits 82.7% below its peak and APT 96.8%. Each figure here can be rechecked on DefiLlama, Tokenomist and the public endpoints named above.

FAQ

?How much does a transaction cost on Sui compared with Aptos?

Dividing DefiLlama's 24-hour chain fees by the measured transaction counts gives roughly $0.000011 per transaction on Sui and $0.0000027 on Aptos, rough derived estimates that include system transactions in the count. Sui's reference gas price read 100 on 5 November 2026 (Sui GraphQL). The Aptos Foundation has also proposed a tenfold gas-fee increase in Aptos Improvement Proposal 140 (AIP-140), which would raise Aptos costs while keeping them fractions of a cent.

?Can a Move contract written for Aptos be deployed on Sui unchanged?

No. Mysten Labs' developer FAQ states that Move on Sui has a different object model, abilities system and standard library, and tells developers not to apply Aptos patterns such as the aptos_framework, signer-based ownership or the move_to and borrow_global operators. A contract has to be redesigned around objects and ownership types, then audited again for the new chain.

?Can Sui validators freeze any wallet they choose?

In the Cetus case, validators independently chose to quarantine transactions from the attacker's addresses, according to Cetus. Returning the funds took a protocol upgrade approved by validators holding 90.9% of stake after a vote in May 2025. The episode shows that a validator majority can act on specific addresses, which is why critics raised decentralization concerns, and why the answer depends on validator coordination rather than a single key.

?How many validators secure Sui and Aptos, and how concentrated is their stake?

Sui's epoch 1302 had 125 active validators holding 7.07 billion SUI of stake, with voting power capped at 3.11% for the largest. Aptos had 83 active validators and 767.6 million APT staked, with the top 14 holding over one third of voting power against 18 on Sui (Sui GraphQL and Aptos fullnode API, 5 November 2026). Staked totals include delegated and locked tokens, so they do not map onto circulating supply.

?Is Mysticeti the same as Narwhal and Bullshark?

No. Narwhal and Bullshark are the earlier data-availability and ordering pair that 2023 and 2024 research described for Sui. Mysten Labs says Mysticeti first launched on mainnet in July 2024 as a new generation of directed-acyclic-graph consensus, and Mysticeti v2 followed in November 2025 with validation folded into consensus.

?What are Walrus and Seal, and do they run on Sui?

Walrus is decentralized storage and Seal is access control, and both launched on mainnet in 2025 as parts of what Sui calls its stack (Sui, 2025). They sit beside the chain rather than inside its consensus. Applications use Walrus for large data that would be costly to hold in on-chain objects, and Seal to govern who may read it.

?What is Decibel on Aptos?

Decibel is a fully on-chain trading engine co-developed by the Decibel Foundation and Aptos Labs that combines spot, perpetuals and margin trading, with every order, match and cancel executed on-chain (Aptos Labs, 2025). DefiLlama listed Decibel at $29.0M of TVL on 5 November 2026, the largest non-exchange protocol on Aptos by that measure.

?How do Sui and Aptos compare with Solana in total value locked?

DefiLlama showed Solana at $5.57B of chain TVL ↗ on 5 November 2026, about 12 times Sui's $462.7M and 136 times Aptos' $40.8M. Ethereum led at $45.82B. The comparison covers DeFi capital only and says nothing about user activity, fees or developer count, which differ again across the three chains.

References / Sources