BNB Chain deployed the BEP-675 upgrade on its BSC Testnet on August 7, pushing throughput from 1,237 transactions per second to 2,324 TPS. That’s an 88% jump, achieved without changing the block interval or gas limit.
How BEP-675 actually works
Before this upgrade, the BSC block-building process had a significant redundancy problem. Block builders would assemble and execute transactions, then validators would re-execute those same transactions to verify them.
BEP-675 introduces a new mechanism called SendBidBlock, which allows block builders to submit fully executed blocks directly. Validators can then skip the redundant re-execution step, trusting the pre-executed results while maintaining the chain’s security model.
The performance gains from removing that redundancy are dramatic. Critical path validator execution time dropped from approximately 125ms to just 15ms. To put that in perspective, the execution step that previously consumed more than a quarter of each 450ms block interval now takes up roughly 3% of it.
That freed-up headroom translates directly into higher gas utilization. Median gas usage per block jumped from 29.49M to 98.99M, meaning blocks that were previously using less than a third of their 100M gas limit are now filling up almost completely. Same block size, same block timing, dramatically more actual computation per block.
The upgrade was first drafted as a proposal on April 10 and went live on testnet roughly four months later. Legacy SendBid flows remain supported for backward compatibility, though builders who want to use the new SendBidBlock mechanism need to operate a full node.
The testing setup and what comes next
BNB Chain ran the testnet evaluation using an internal cross-region QANet setup designed to mirror the actual mainnet topology. By simulating cross-region conditions, the 2,324 TPS figure should be a closer approximation of what mainnet could actually deliver.
The testing covered various transaction workloads rather than just simple token transfers.
BEP-675 sits within a broader H2 2026 technical roadmap for BNB Chain. The chain has been on an aggressive scaling trajectory, having already reduced block intervals to 450ms and pushed benchmark throughput close to 5,200 TPS in earlier phases during 2025 and early 2026. The next target is another doubling of mainnet throughput, with a longer-term goal of achieving 10x improvements over current baseline performance.
Following the successful testnet phase, the immediate next steps include mainnet-scale validation. The roadmap also calls for additional enhancements including FOCIL (which relates to forced inclusion lists, a mechanism designed to prevent censorship at the block production level) and Block-Level Access Lists, which could further optimize execution efficiency.
Why MEV matters here
BEP-675 reduces the operational overhead that MEV infrastructure imposes on the chain’s critical path. BNB Chain explicitly framed the upgrade as addressing bottlenecks caused by MEV inefficiencies. By redesigning the submission mechanism so that builders deliver fully executed blocks, the redundant re-execution step that was partly a consequence of trust assumptions baked into MEV-aware architectures is eliminated.
If the mainnet deployment matches testnet results, BNB Chain will have nearly doubled its practical throughput without requiring users or dApp developers to change anything about how they interact with the network. Finality guarantees and block timing remain identical.