BNB Chain (BNB) switched on its Pasteur hard fork Tuesday, closing two validator loopholes and lifting testnet throughput 88% to 2,324 transactions per second without slowing blocks.
Key Points:
- Pasteur went live on BNB Smart Chain at 02:30 UTC on Aug. 25, bundling three proposals under the BEP-673 package.
- BEP-682 rejects duplicate validator entries in bridge checks, while BEP-695 strips authority from rotated validator keys.
- Internal testing raised throughput from 1,237 to 2,324 transactions per second at an unchanged 450-millisecond block interval.
Pasteur Hard Fork Activation
BNB Chain confirmed that Pasteur reached the BNB Smart Chain mainnet at 02:30 UTC on Tuesday, roughly five weeks after the code began running on the public testnet. Every mainnet node had to move to client version v1.7.7 beforehand, or drop out of sync the moment the fork switched on at the scheduled block.
The fork bundles three proposals under the BEP-673 package, two of them aimed squarely at validator power. BEP-682 targets the precompile that counts validator signatures on cross-chain transfers, where a crafted validator set could previously list the same signer more than once and clear the approval threshold. BEP-695 strips admin authority from rotated consensus keys, carries pending slashing evictions across to the new key and bars blacklisted addresses from signature-based governance votes.
Also Read: What Happens To Bitcoin When Washington Declares Economic War On Iran?
BEP-675 Throughput Test Results
The third proposal changes how blocks travel from the specialist builders that assemble most of them to the validators that sign them, and it cuts out execution work the two sides used to repeat. Builders can now hand over a block they have already executed, leaving the validator to check it against consensus rules, sign and broadcast it, then finish verification afterward.
Testing on QANet, an internal network built to mirror the geographic spread of mainnet validators, trimmed the validator's share of the critical path from 125 milliseconds to 15. Average gas used per block climbed from 46.35 million to 84.15 million against an unchanged 100 million ceiling. Finality lag held steady.
The team cautioned that those figures came from a controlled test workload rather than live mainnet traffic, and it made no promise the same gains would show up once real users hit the chain.
BSC Block Time Upgrades
Pasteur breaks from the two upgrades before it, both of which chased lower latency rather than fuller blocks. Fermi cut block times to 0.45 seconds in January, and the Osaka and Mendel fork steadied the network at that pace in April.
Those releases reshaped the first half of 2026 for BSC. The chain trimmed its block interval to 450 milliseconds, brought in-memory finality down to roughly 650 milliseconds and nearly doubled benchmark throughput to about 5,200 transactions per second.
Pasteur is the first fork of the year aimed at filling the blocks the network already produces instead of squeezing more of them into a minute.
Read Next: Eric Trump Denies New Token Launch Rumor With 3 Trump Tokens Far Off Peaks
