Glossary · Blockchain basics

Block time

Published 27 Apr 2025 · Updated 21 Oct 2026 · 12 min read

Block time is the average interval between consecutive blocks on a blockchain. Bitcoin targets ten minutes through proof of work, with random gaps. Scheduled chains follow fixed intervals, with Ethereum at 12 seconds and Solana at 200 milliseconds since October 2026. Finality is a separate, later step that protects blocks against reversal.

  • Bitcoin targets a 10-minute average, but proof-of-work block times are random.
  • Ethereum has 12-second slots, and Solana's target has been 200 milliseconds since 9 October 2026.
  • Block time describes cadence, not finality or transaction capacity per second.
How block time is setBlock times across blockchainsWhy Bitcoin block times varyHow Bitcoin keeps ten-minute blocksWhen block times changedWhat shorter block times costBlock time in contracts and appsDesigning around block timeBlock time FAQ

How block time is set

A target is a protocol's intended interval. A measured average describes the blocks that actually appeared during a stated period. Scheduled slots can remain empty. A block exists only when one is produced. Network rules therefore need different readings, even when explorers present every value in seconds. A published average also depends on the measurement window and timestamp source.

Proof of work

Bitcoin targets 600 seconds. Its mining difficulty changes every 2,016 blocks to keep the long-run average near that target. Before 15 September 2022, Ethereum's proof-of-work target was about 13.3 seconds per block.

Proof of stake

Since the Merge on 15 September 2022, Ethereum has 12-second slots, with 32 per epoch. An offline proposer can leave one empty.

BFT rounds and leaders

Byzantine fault-tolerant rounds use named producers and votes. Solana assigns four consecutive slots per leader, BNB Smart Chain uses turns and Monad votes for finalization.

Rollup sequencers

OP Mainnet and Base have 2-second blocks. Base streams ten 200-millisecond Flashblocks as preconfirmations. Arbitrum's sequencer produces blocks only when transactions need sequencing, at most every 250 milliseconds.

Not every distributed ledger has a single shared chain. The Canton Network explainer describes ordering within each synchronizer. Its sequencer orders transactions for that synchronizer, and nodes store only their parties' data. Without one global chain replicated by every node, Canton has no network-wide block time.

Block times across blockchains

Inclusion puts a transaction into a block. Transaction confirmations count that block and each one above it, while finality describes protection against reversal. The three proof-of-work markers use six blocks deep for comparison, not a shared deposit rule. Their assurance remains probabilistic. Targets and measured averages appear separately, since an observed interval is not a guaranteed schedule. The "Final" label indicates that a network has passed its marker.

Figure 1 · Nine networks side by sideBlocks within a time windowMove the window from one second to two hours to compare block counts and each network's assurance marker.
1 minute
Bitcoin10 min target, 9 min 56 s measured0.10 blocks6 blocks, about 1 h
Litecoin2.5 min target, 148 s measured0.40 blocks6 blocks, 15 min
Dogecoin1 min target, 63 s measured1.0 block6 blocks, 6 min
Ethereum12 s slots, 12.04 s measured5.0 blocksFinality, 2 to 3 epochs
BNB Smart Chain0.45 s since January 2026133 blocksFinal
Polygon PoS1.5 s measured40 blocksFinal
OP Mainnet2 s blocks30 blocksFinal in 15 to 30 min
Monad0.3 s since July 2026200 blocksFinal
Solana0.2 s slots since October 2026300 blocksFinal

Across several minutes, more networks pass their markers. Bitcoin reaches six blocks deep in about an hour.

Targets follow network documentation. Page team averages date to 21 October 2026. Proof-of-work markers show six blocks deep.

NetworkBlock or slot timeFinalityHow it is set
BitcoinBlock or slot time600 s target, 596.2 s measured from 4 June to 21 October 2026FinalitySix blocks, about an hour, probabilistic assuranceHow it is setDifficulty retarget every 2,016 blocks
LitecoinBlock or slot time150 s target, 148.3 s measured over 5,760 blocks to 21 October 2026FinalitySix blocks, about 15 minutes, probabilistic assuranceHow it is setDifficulty retarget every 2,016 blocks, about 3.5 days
DogecoinBlock or slot time60 s target, 63.3 s measured over 24 hours on 21 October 2026FinalitySix blocks, about six minutes, probabilistic assuranceHow it is setDigiShield difficulty adjustment after every block
EthereumBlock or slot time12 s slots, 12.035 s measured over 7,200 blocks to 21 October 2026FinalityTwo to three epochs, about 13 to 19 minutesHow it is setFixed 12 s slots, 32 slots per epoch
BNB Smart ChainBlock or slot time0.45 s since 14 January 2026, 0.4502 s measured on 21 October 2026FinalityAbout 1.125 s to finality, BEP-619 estimateHow it is set16 consecutive blocks per validator turn since June 2025
Polygon PoSBlock or slot time1.5 s measured over 20,000 blocks on 21 October 2026FinalityDeterministic finality in 2 to 5 seconds with Heimdall v2How it is setOne producer per span, from a small eligible set
OP MainnetBlock or slot time2 s blocks, per the Optimism docsFinalitySafe after a few minutes, finalized after about 15 to 30 minutesHow it is setThe sequencer builds a block every two seconds
MonadBlock or slot time300 ms blocks since MIP-12 on 23 July 2026Finality600 ms finality since 23 July 2026How it is setMonadBFT vote pace, cut to 300 ms by MIP-12
SolanaBlock or slot time200 ms target, 212 ms per slot measured over one hour on 21 October 2026FinalityAbout 150 ms target with Alpenglow, live since 21 October 2026How it is setStake-weighted leader schedule, four consecutive slots per producer

Why Bitcoin block times vary

Mining is a random search, not a countdown. With constant hashrate, Bitcoin's gaps follow an exponential distribution around a 10-minute mean. The model puts 36.8% beyond 10 minutes, 13.5% beyond 20 and 5.0% beyond 30. Only 0.25% exceed an hour, while 9.5% finish within a minute. The median is 6.9 minutes. Starting at a random moment still gives a 10-minute average wait, because long gaps occupy more of the clock.

Across 20,160 blocks from 4 June to 21 October 2026, the median gap was 7.0 minutes. Of those gaps, 36.8% lasted at least 10 minutes, and 61 lasted an hour or more. The longest reached 112 minutes. Results fit a memoryless process, with past waiting not making the next block due. Miners set timestamps, and 585 blocks were stamped earlier than their parents. Each timestamp must exceed the median of the previous 11 blocks, not necessarily the parent's time.

Figure 2 · Bitcoin, 20,160 blocksRandom gaps, steady averageCompare the histogram with its model, adjust the elapsed wait or switch to one day to see uneven arrivals.
View
Real gaps between blocks, by minuteExpected at a 10-minute average
Blocks 972,344 to 972,488, 20 to 21 October 2026, times in UTC
0 minutes
Gaps under 1 minute10.7%Includes 585 blocks stamped earlier than their parent. The model expects 9.5%
A block within 10 more minutes62.1%Real blocks after waiting 0 min. The model says 63.2% at any point
Average wait from a random moment10.1 minNot 5 minutes. Long gaps cover more of the clock

The 140-day histogram follows the exponential curve, with many short gaps and a long tail.

Data cover 20,160 blocks from 4 June to 21 October 2026, sourced from mempool.space. Miners set timestamps.

How Bitcoin keeps ten-minute blocks

Every 2,016 blocks, Bitcoin compares elapsed time with 1,209,600 seconds, or two weeks. It scales difficulty to bring the average back toward ten minutes. Each adjustment is capped at four times up or down. The code measures 2,015 intervals instead of 2,016, an off-by-one that enables the timewarp attack. In October 2026, BIP-54 proposes timestamp rules to close this gap, but has no activation date. The 20 complete periods that began in 2026 average 602.5 seconds.

Figure 3 · Every Bitcoin retarget since 2010Retargets across changing hashrateSelect a period since 2010, then change hashrate and its timing to compare the average, duration and next adjustment.
Average block time per 2,016-block period
13 June 2021
Blocks 687,456 to 689,471, 13 June 2021 to 3 July 2021: 13.9 minutes per block on average. The next retarget changed difficulty by -27.94%.
What happens when hashrate changes
Scenario
-50%
60% through
Average block time this period14.0 minTarget 10 minutes
Length of the 2,016-block period19.6 daysTarget 14 days
Next difficulty change-28.6%New difficulty restores 10-minute blocks

When hashrate falls, blocks slow until difficulty falls. The opening June 2021 period illustrates a historical slowdown.

Historical periods come from mempool.space on 21 October 2026. The simulator uses a simplified model.

The period starting at block 971,712 on 16 October 2026 averages 9.4 minutes through 21 October 2026. The next retarget, at block 973,728, is estimated around 29 October 2026, with difficulty rising about 6%. Bitcoin halvings occur every 210,000 blocks. Block 840,000 arrived on 20 April 2024. The 21 October 2026 estimate puts block 1,050,000 in spring 2028.

When block times changed

Intervals change for different reasons. A mining adjustment reacts to completed work, while an upgrade can change the production schedule itself. These examples separate automatic corrections from deliberate protocol changes, including changes to how quickly blocks become irreversible.

  1. October 2017

    The bomb and Byzantium

    Ethereum's difficulty bomb raised mining requirements on a schedule. Around Byzantium on 16 October 2017, 10,000-block samples averaged 29.70 seconds before the upgrade and 15.41 seconds afterward. Delays in February 2019, January 2020 and June 2022 repeated that pattern.

  2. July 2021

    Bitcoin's largest downward retarget

    After China's mining ban, the period beginning 13 June 2021 averaged 13.9 minutes per block. The 3 July 2021 retarget at block 689,472 cut difficulty by 27.94%, the largest downward adjustment on record.

  3. September 2022

    Ethereum changes consensus

    Ethereum replaced proof of work with proof of stake through the Merge on 15 September 2022. The final 10,000 mining blocks averaged 14.51 seconds, compared with 12.19 seconds for the first 10,000 under the new schedule.

  4. November 2025

    Avalanche adopts dynamic delays

    Granite on 19 November 2025 replaced Avalanche C-Chain's old 2-second target with a dynamic minimum delay set by validators. Blocks average 0.90 seconds over 20,000 blocks on 21 October 2026.

  5. January 2026

    BNB Smart Chain cuts again

    Fermi on 14 January 2026 set a 0.45-second interval. It followed Lorentz's 1.5 seconds on 29 April 2025 and Maxwell's 0.75 seconds on 30 June 2025. The sequence began with a 3-second interval.

  6. July 2026

    Monad shortens both waits

    The 23 July 2026 MIP-12 upgrade cut the interval from 400 to 300 milliseconds. Finalization now takes 600 milliseconds rather than 800.

  7. October 2026

    Solana changes cadence and consensus

    Solana reached 200-millisecond slots on 9 October 2026, the last of four reductions. On 21 October 2026 Alpenglow replaced TowerBFT, targeting finality of about 150 milliseconds. This is an expectation, not a constant delay.

What shorter block times cost

Faster cadence leaves less room for information to travel. The engineering question is how production, propagation and voting fit within the available interval. Capacity is a separate parameter, so a lower waiting time does not establish a higher transaction rate.

Propagation and stale blocks

Stale block risk is roughly t/(T+t), using transit time t and interval T. Bitcoin's measured fork rate in 2012 was 1.69%. Ethereum's first 280,000 blocks in 2015 had a 7.41% uncle block rate.

Less time for votes

Fast chains name producers ahead and finalize through votes. BEP-619 warns that reducing intervals also reduces the time available for vote propagation.

Timing games

Ethereum proposers can delay publication to earn more MEV within their allotted slot. Researchers warn of missed slots, reorgs and incorrect attestations.

Throughput remains separate

BNB Smart Chain's 2025 cuts shortened the interval more than they lowered gas limits, so capacity per second rose. Solana's SIMD-0525 halved slot time but scaled per-slot limits down, keeping throughput about the same.

Block time in contracts and apps

Ethereum's block.timestamp follows genesis time plus 12 seconds per slot. block.number counts produced blocks, so it falls behind the slot count after missed slots. Neither timestamps nor blockhash provide dependable randomness. Compound's Governor Bravo assumes 15-second blocks for its 5,760-block minimum voting period. At 12 seconds, that becomes 19.2 hours rather than 24. A block-based governance window can therefore drift after a network upgrade. OpenZeppelin supports timestamp clocks through ERC-6372 from version 4.9, avoiding dependence on changing block intervals.

Arbitrum's block.number returns the block of its first non-Arbitrum ancestor chain. ArbSys(100).arbBlockNumber() gives the Arbitrum count. BNB Smart Chain's January 2026 cut also moved block-based timing. Since Solana's 9 October 2026 change, fixed slot-to-time conversions drift, and a blockhash expires after 150 blocks, about 30 seconds. Solana suggests getBlockTime. Bitcoin's time-based locks use the median of the last 11 blocks, about an hour behind wall-clock time. Relative time locks count blocks or 512-second units. New coins need 100 blocks, about 16.7 hours.

Designing around block time

A timing assumption can reach from a contract into deposit handling, indexing and monitoring. For blockchain development projects, PixelPlex considers five connected areas. The goal is to make the product's behavior explicit when blocks arrive early, arrive late or stop matching assumptions made during development.

PixelPlex case icon

PixelPlex has worked on both sides of block time. It built the Echo protocol, where a new committee is chosen at random for each block. It worked on the Qtum platform, whose governance can change block time. For the mining client PEGA Pool, it built the PegaPool mining pool and modified Antminer S19 firmware for the miners who produce blocks.

  1. Contract clocks

    We identify whether each timelock and deadline depends on elapsed time or block counts, including governance windows and transaction expiry.

  2. Deposit assurance

    We separate inclusion, confirmation counts and finality in deposit rules, so an interface does not present them as interchangeable states.

  3. Indexing and reorgs

    We account for reorgs and missed slots when describing indexer behavior, rather than assuming uninterrupted block production.

  4. Chain parameters

    For a chain with its own rules, we connect the production interval to capacity, voting time and parameter governance.

  5. Production monitoring

    We distinguish a target from observed intervals in monitoring, including changes in hashrate and production gaps that need interpretation.

Discuss the timing assumptions in your product

We can discuss your network assumptions and where waiting, block counts or irreversible settlement affect the product's behavior.

Discuss your project

Block time FAQ

What is Bitcoin's target block time?

Bitcoin aims for one block every ten minutes, not a precise ten-minute pause between blocks. Across 20,160 blocks from 4 June to 21 October 2026, the measured average is 596.2 seconds. That sample spans heights 952,328 to 972,488. A displayed average describes a sample, while the target stays fixed.

What is Ethereum's block time after the Merge?

Ethereum follows a 12-second schedule, with empty slots making measured gaps slightly longer. EIP-7782's six-second proposal remains a draft in October 2026. EIP-8198's ten-second proposal is considered for inclusion in Hegotá, not scheduled. Glamsterdam keeps the existing interval and has no mainnet date in October 2026.

What is Solana's block time in October 2026?

Solana has targeted 200-millisecond slots since 9 October 2026, with a 212-millisecond measured average during one hour on 21 October 2026. That measure tracks slot timing, and a slot can pass without a block. An epoch contains 432,000 slots, about 24 hours at the target. The first full epoch at 200 milliseconds averaged 218.3 milliseconds.

Why do Bitcoin blocks arrive at irregular intervals?

A miner's search has no progress bar that makes success inevitable after ten minutes. Under constant hashrate, the model gives roughly a 63% chance of finding a block within the next ten minutes. That chance does not rise simply because the previous wait was long. Changing hashrate can also shift the average between retargets.

How does block time differ from finality?

Block time measures spacing between blocks. Finality means a block cannot be reversed without breaking the protocol's rules. Ethereum illustrates the difference with a 12-second production schedule and roughly 13 to 19 minutes before finalization. Inclusion alone does not reach that later state. Proof-of-work chains instead provide probabilistic assurance, with each later block making reversal less likely.

How many confirmations does a transaction need?

The receiving service sets its own threshold, rather than inheriting one universal number from the network. In October 2026, Kraken credits Bitcoin deposits after three confirmations. At Bitcoin's target, those three blocks represent about 30 minutes, but random arrivals mean the actual wait can differ. For comparison, 64 Ethereum blocks span 12.8 minutes at the scheduled rate.

Do shorter block times make a blockchain faster?

They reduce the interval between chances for inclusion, not automatically the amount of work completed per second. More frequent blocks raise throughput only if each retains its capacity. Solana's October 2026 schedule change illustrates the distinction, with proportionally smaller limits keeping throughput about the same.

Related terms

All definitions
Sources (46)

Get updates about blockchain, technologies and our company

We will process the personal data you provide in accordance with our Privacy policy. You can unsubscribe or change your preferences at any time by clicking the link in any email.

Follow us on social networks and don't miss the latest tech news

  • facebook
  • X
  • linkedin
  • instagram
Stay tuned and add value to your feed