Market cap $2.20 T −3.48 % over 24 hours24h volume $73.58 BBTC dominance 58.1 %ETH 10.0 %Fear & Greed 10 Extreme fearCoins 21,490

Bitcoin developers start Testnet5, a test network where miners can rehearse the four “consensus cleanup” rules; nothing is scheduled for Bitcoin itself

Bitcoin Testnet5 started on 14 November and enforces the BIP 54 consensus cleanup from block 1. No Core release includes it, and mainnet is unchanged.

10 min read

Abstract isometric blocks with a pickaxe, a gear and connected network nodes
Table of contents
  1. Testnet5 started on 14 November, timed to a mainnet block
  2. What changes compared with Testnet4
  3. The four BIP 54 rules Testnet5 enforces
  4. Why miners need a testnet when signet already has the rules
  5. No Bitcoin Core release includes Testnet5 yet
  6. Nothing is scheduled for Bitcoin itself

Bitcoin developers have started Testnet5, a new public test network whose first block was mined on Saturday 14 November 2026 at 05:03 UTC. Its specification, BIP 95 by Pol Espinasa and Fabian Jahr, has carried the status “Deployed” since 17 November. The network replaces Testnet4 and, unlike it, enforces the four rules of the “consensus cleanup” proposal, BIP 54, from block 1.

That is a change to a test network, not to Bitcoin. BIP 54 has no activation date and no activation mechanism on the real network, and the software that follows Testnet5 is not in any Bitcoin Core release yet. What Testnet5 gives miners and blockchain that runs exactly as written when it is called, holding and moving funds according to its code, without an intermediary able to change the outcome.">smart contract holding reserves of two or more tokens, supplied by users, that a decentralized exchange uses to execute trades automatically at a price set by a formula.">pool operators is a place to check, with real hardware and coins that are meant to have no value, that the blocks their software builds would still be valid if the cleanup rules were ever adopted.

Testnet5 started on 14 November, timed to a mainnet block

The start was announced in advance. On 9 November, developer Anthony Towns wrote to the Bitcoin-Dev mailing list that he proposed to mine the genesis block at the end of the week, aligned with the next mainnet difficulty adjustment at block 975,744, and pointed volunteers to a solo-mining pool with a 0% fee. On 13 November Fabian Jahr repeated the invitation on GitHub, about 15 hours before the expected block.

The genesis block carries the timestamp 1794632593, which is 14 November 2026 at 05:03:13 UTC. Its coinbase message contains the date, the height 975,744 and the hash of that mainnet block. BIP 95 calls this an “anti-pre-mine commitment”: nobody could have mined Testnet5 blocks in secret before that mainnet block existed, in the same way the newspaper headline in Bitcoin’s own first block dates it. The timestamp, chosen by the mining software, ended up slightly later than the mainnet block’s, the document adds. The block’s single output is an unspendable OP_RETURN. Block 975,744 opened Bitcoin’s current difficulty period: it is the block of the 9.13% mining difficulty cut of 14 November.

Towns filed the genesis parameters with the BIPs repository at 06:55 UTC the same morning. They were merged on 17 November at 10:19 UTC, which is when BIP 95 reached version 1.0.0. The chain is five days old; its paperwork, two. The idea is older: the first draft of the BIP is dated 2 June 2026 and it received its number on 22 June.

What changes compared with Testnet4

Testnet4, specified in BIP 94 in 2024, kept an old testnet convenience: if 20 minutes passed without a block, the next one could be mined at the minimum difficulty, so that someone with an ordinary computer could still obtain test coins. BIP 95 says the exception was “systematically and persistently exploited” soon after Testnet4 began. The result, according to the document, was constant reorganisations of a few blocks as several minimum-difficulty blocks competed for the tip of the chain, on a network people use to test software.

Testnet5 removes the exception. Its proof-of-work rules are those of mainnet, with a floor: the minimum difficulty is about 1,000,000, against 1 on Testnet4. The authors write that this level slows the rush of blocks right after a launch while staying low enough for “a handful of at-home miners or a single ASIC” to keep the network usable. Our guide to hashrate and the difficulty adjustment explains how that number moves with the computing power on a network.

Abstract hourglass and stacked blocks beside a computer chip and a rising arrow
Testnet5 drops the 20-minute exception and sets a minimum difficulty of about 1,000,000.
Testnet4 (BIP 94)Testnet5 (BIP 95)
BIP authorsFabian JahrPol Espinasa, Fabian Jahr
Minimum difficulty1About 1,000,000
20-minute minimum-difficulty exceptionYes, in a modified formNo
BIP 54 consensus cleanup rulesNot enforcedEnforced from block 1
Timestamp rule against timewarp600-second margin7,200-second margin, as in BIP 54
Default peer-to-peer port4833318353

Why not repair Testnet4? BIP 95 answers that any real fix would have been a hard fork of a network whose attackers could disrupt the transition, and that a patch would only have pushed Testnet4 further from mainnet. A fresh chain was cheaper. The two networks are incompatible, with separate genesis blocks and separate sets of coins, although an address has the same format on both.

Testnet coins are not bitcoin. They live on a separate chain and cannot be moved to mainnet. They are also supposed to be free: BIP 94 quotes developer Jameson Lopp on “the fundamental principle of testnet coins having no value”, which he considered broken on Testnet3 once its coins were being bought and sold. Testnet5’s genesis block follows that principle. Its output cannot be spent, and its message rules out blocks mined in advance.

The four BIP 54 rules Testnet5 enforces

BIP 54, “Consensus Cleanup”, is written by Antoine Poinsot and Matt Corallo. It builds on an earlier proposal by Corallo, a draft that Bitcoin Optech dates to 2019 and describes as deferred before work resumed in 2023. It was assigned its number on 11 April 2025 and reached version 1.0.0, with the status “Complete”, on 22 May 2026. It is a soft fork proposal: it only tightens the rules, so a block valid under BIP 54 is also valid for a node that has never heard of it. It bundles four fixes for weaknesses that have been known for years.

The first is the timewarp attack. Bitcoin recalculates its mining difficulty every 2,016 blocks from block timestamps, and a quirk in that calculation lets miners who control a majority of the hashrate lie about timestamps to push the difficulty down. The BIP says an attacker could bring it to its minimum within 38 days and produce blocks, with their newly created coins, far faster than one every ten minutes. The fix constrains two timestamps in each period: the first block may not be dated more than 7,200 seconds (two hours) before the block preceding it, and the last block may not be dated earlier than the first.

The second concerns blocks that are slow to verify. A block built for the purpose can take several minutes to validate on a high-end machine and up to a few hours on a weak one, the BIP says, which a miner could use against competitors. The proposed limit is 2,500 signature operations per transaction, counted in the scripts of its inputs whether or not they are executed, for every transaction except the coinbase. According to the authors it cuts the worst case by a factor of 40, and 2,500 is the tightest figure that leaves every normal standard transaction valid.

Third, 64-byte transactions become invalid. A transaction of exactly 64 bytes (without witness data) is the same size as an inner node of the Merkle tree that summarises a block, and that coincidence makes it possible to give a light wallet a false proof that a payment is in a block. Such transactions have been non-standard since 2019 and none has been used since 2016, according to the BIP. The point is still debated: the BIP itself records that some developers consider the gain too small to justify outlawing a transaction size.

Fourth, every coinbase transaction must be unique. The coinbase is the transaction that pays the miner. BIP 54 requires its lock-time field to equal the block height minus one, and its sequence field not to be at its maximum value, which makes two identical coinbases impossible. Without it, nodes would have to resume an old and costly duplicate check from block 1,983,702, about a million blocks from now.

RuleWeakness it closesWho has to act
Timestamp limits at the edges of each 2,016-block periodTimewarp: a mining majority lowers the difficultyBlock templates from Bitcoin Core 29.0 and later already comply
At most 2,500 legacy signature operations per transactionBlocks that take minutes or hours to verifyTemplates from Bitcoin Core 30.0 and later already comply
No transaction of exactly 64 bytesFalse inclusion proofs shown to light walletsNot relayed or mined by Bitcoin Core since 0.16.1
Coinbase lock time equal to height minus 1Duplicate coinbase transactionsMining pool software, which builds the coinbase
Abstract shield and magnifying glass next to a pickaxe and a document
Three of the four BIP 54 rules are already respected by Bitcoin Core templates; the coinbase rule is for pool software.

Why miners need a testnet when signet already has the rules

The last line of that table is the reason Testnet5 exists in this form. Bitcoin Core already avoids producing templates that would break the first three rules. The coinbase is another matter: it is “usually crafted by mining pool software”, BIP 54 notes, and to the authors’ knowledge there is no open-source reference for such software in broad use. Each pool has to check its own code.

The cleanup rules already run on signet, another test network, through the Bitcoin Inquisition client. But signet, defined in BIP 325, requires each block to carry a signature that satisfies the network’s challenge script, on top of proof of work. Blocks come from whoever holds the signing keys, not from open competition between miners. BIP 95 draws the consequence: signet “does not allow miners to test that their software reliably follows the rules of BIP 54”. On Testnet5 a pool can point real machines at the chain, and a block with a wrong coinbase lock time or a bad timestamp is rejected by the other nodes, exactly as it would be on a mainnet with BIP 54 active. A mistake there costs nothing. Readers new to the subject will find templates, pools and coinbases explained in our guide to how bitcoin mining works.

No Bitcoin Core release includes Testnet5 yet

Running a Testnet5 node today means compiling code that has not been merged. The reference implementation, Bitcoin Core pull request #35861 by Fabian Jahr, was opened on 1 August and was still open, unmerged, at 05:48 UTC on 19 November, with its last update on 14 November. It rests on a second pull request, #35793 by Antoine Poinsot, also open and last updated at 05:08 UTC on 19 November. Its title states its scope: “Implement BIP 54 (Consensus Cleanup) without mainnet activation”. Its description says the rules are switched on only in regtest, the private mode developers run on their own machines, and adds: “Mainnet activation, if any, is to be considered separately.”

The latest release, Bitcoin Core 32.1 of 28 October, knows nothing of Testnet5. Jahr’s plan, written in the pull request, is to settle the code in review first: “We can then ship a patched v32 release with Testnet 5 included.” No date is given. For now the network has no DNS seeds, so a node finds its peers only if the operator types in the address of one by hand, and wallet support is at the same stage: a pull request adding Testnet5 to Sparrow Wallet was opened on 17 November and has not been merged.

Nothing is scheduled for Bitcoin itself

The proposal describes rules. It contains no deployment section, no block height and no date. Activating a soft fork on Bitcoin takes a separate decision on the method and broad agreement among developers, miners, node operators and businesses, a process described in our guide to how Bitcoin Improvement Proposals are adopted. None of that has begun for the consensus cleanup.

The only timeframe on the public record is one developer’s opinion. In a message to the mailing list on 2 October, Antoine Riard wrote that it might be realistic to start thinking about an activation in the second half of 2027 if community consensus is there by then, and added: “I’m only speaking for myself.” He is not an author of the BIP.

For someone who holds or uses bitcoin, nothing changes and there is nothing to do. The next steps are on GitHub: the merge of pull request #35793, then of #35861, then the patched 32.x release that Jahr has announced without a date.

Sources

  1. BIP 95: Testnet 5 (Pol Espinasa, Fabian Jahr), version 1.0.0 of 17 November 2026 github.com
  2. BIP 54: Consensus Cleanup (Antoine Poinsot, Matt Corallo) github.com
  3. BIP 94: Testnet 4 (Fabian Jahr) github.com
  4. Bitcoin Core pull request #35861: Testnet 5 (BIP95) github.com
  5. Anthony Towns, "Testnet 5 launch", Bitcoin-Dev mailing list, 9 November 2026 gnusha.org
  6. Antoine Riard, "Getting BIP54 past the finish line", Bitcoin-Dev mailing list, 2 October 2026 groups.google.com

Leave a Reply

Your email address will not be published. Required fields are marked *