BlockRazor vs bloXroute: A Comparison of Low-Latency Infrastructure Capabilities

Benchmark conclusion: This article compares BlockRazor and bloXroute under consistent test conditions in Virginia, Frankfurt, Dublin, and Tokyo across Node Stream, Block Stream, mempool transaction monitoring, and public transaction submission. Results are reported using first arrive share, observed transaction coverage, transaction-race win rate, and same-block index position. BlockRazor led every reported regional comparison:

• Node Stream: BlockRazor's first arrive share ranged from 53.72% to 75.31%, compared with 17.78% to 45.18% for bloXroute.
• Block Stream: BlockRazor delivered 65.38% to 95.77% of observed blocks first, versus 4.23% to 34.62% for bloXroute.
• Mempool transaction monitoring: BlockRazor's first arrive share ranged from 56.71% to 85.09%, while its observed transaction coverage reached 99.46%–99.76%; bloXroute recorded 14.91%–43.29% first arrive and 69.31%–78.44% observed coverage.
• Public transaction submission: BlockRazor won 747 of 797 races, an overall win rate of 93.73% and more than 90% in every region. Among same-block pairs, it secured the earlier index in 91.62%–96.30% of each region's samples.

Benchmark Scope and Metrics

This benchmark focuses on low-latency BSC data delivery and public transaction submission. The four test categories use the following decision rules:

Node Stream Performance

Node Stream accelerates the synchronization of the latest blocks and world state to a local node, reducing the window in which that node operates on stale state. For quantitative teams, trading bots, searchers, and node engineering teams that depend on local nodes for strategy decisions, transaction preparation, or on-chain analysis, faster state synchronization provides a more timely foundation for computation, risk control, and execution.

First arrive share across four regions

First arrive identifies which source delivered a block to the test client first. It measures arrival frequency: a higher share means that a provider was the earliest source for a larger proportion of observed blocks.

BlockRazor recorded a higher first arrive share than bloXroute in all four regions. Virginia measured 75.31% versus 17.78%, a difference of 57.53 percentage points. Frankfurt measured 74.68% versus 24.38%, a difference of 50.30 points. Tokyo measured 68.40% versus 30.64%, a difference of 37.76 points. Dublin was the closest at 53.72% versus 45.18%, an 8.54-point difference. Other P2P sources accounted for less than 7% in every region.

How far ahead was BlockRazor when it won?

The following table includes only blocks where BlockRazor arrived first and shows the conditional lead-time distribution for those wins. These values measure BlockRazor's winning margin, not end-to-end service latency.

Among the samples where BlockRazor arrived first, Frankfurt had the highest P50 lead at 6.295 ms. Dublin showed the largest lead at the higher percentiles, with P90, P95, and P99 values of 14.491 ms, 19.973 ms, and 25.798 ms. Tokyo recorded a P50 lead of 3.928 ms and a P99 lead of 24.512 ms. In other words, BlockRazor maintained millisecond-level median and higher-percentile advantages within the samples it won.

Block Stream Monitoring

Block Stream provides low-latency delivery of the latest blocks and their confirmed transactions, allowing monitoring and strategy systems to update confirmed state sooner. It can support monitoring for target addresses, popular contracts, and anomalous transactions, while also supplying block headers, transaction lists, and subsequent block-producer information for block-level analysis. Combined with mempool and transaction-submission services, it forms a monitoring, decision, and execution workflow. This section compares which provider delivered each observed block first.

Virginia recorded the highest BlockRazor first arrive share at 95.77%, versus 4.23% for bloXroute. Tokyo measured 80.63% versus 19.37%, a difference of 61.26 percentage points and 38,434 versus 9,234 first arrive blocks. Frankfurt measured 77.07% versus 22.93%, while Dublin measured 65.38% versus 34.62%. BlockRazor's first arrive share ranged from 65.38% to 95.77% across the four regions, although the size of the lead varied substantially by region.

Mempool Transaction Monitoring

Mempool transaction monitoring provides low-latency access to publicly propagated pending-transaction signals before on-chain inclusion, allowing strategies to begin evaluation, parameter calculation, and risk control sooner. It can support pending-transaction monitoring, Smart Money tracking, backrunning, copy trading, and sniping. This section evaluates two dimensions: which provider arrived first for transactions observed by both services, and what share of all unique transaction hashes each provider observed.

For first arrive, BlockRazor measured 85.09% in Virginia, 76.92% in Tokyo, 61.10% in Frankfurt, and 56.71% in Dublin, leading bloXroute in all four regions. For observed share, BlockRazor ranged from 99.46% to 99.76% across the four regions, compared with 69.31% to 78.44% for bloXroute. Dublin had the largest observed-share gap at 30.45 percentage points. Here, observed share is the number of unique transaction hashes seen by a provider divided by the union of unique hashes seen by both providers during the test; it does not measure on-chain inclusion.

Public Transaction Submission

Public transaction submission sends signed transactions into public propagation and the on-chain execution path through a low-latency route. It can serve as a fast submission entry point without requiring an additional tip in the transaction. This approach suits teams prioritizing inclusion speed without making private propagation or MEV protection the primary requirement. The benchmark used paired submissions: the earlier inclusion block determined the winner, and if both transactions landed in the same block, the lower transaction index determined the winner.

Across 797 valid public-transaction races, BlockRazor won 747, for an overall win rate of 93.73%. In Virginia, it won 192 of 199 pairs, or 96.48%. In Tokyo, it won 188 of 198 pairs, or 94.95%. It won 185 of 200 pairs in Dublin, or 92.50%, and 182 of 200 pairs in Frankfurt, or 91.00%. BlockRazor's win rate exceeded 90% in all four regions.

Same-block index comparison for public transactions

When both transactions landed in the same block, the lower index occupied the earlier position. The following table includes only same-block pairs. Average and maximum index advantages are calculated separately from the pairs won by each provider.

Many public-transaction comparisons were decided by ordering inside the same block. BlockRazor secured the earlier index in 182, 175, 173, and 178 same-block pairs in Virginia, Frankfurt, Dublin, and Tokyo, respectively, representing 96.30%, 91.62%, 92.02%, and 94.68% of each region's same-block sample.

Final Conclusion

Under the test conditions and decision rules used in this article, BlockRazor led the Node Stream, Block Stream, and mempool transaction first arrive comparisons in Virginia, Frankfurt, Dublin, and Tokyo; its mempool observed share remained between 99.46% and 99.76%, and its public transaction-race win rate exceeded 90% in every region. These results show a consistent BlockRazor advantage across the BSC low-latency data-delivery and public transaction-submission metrics covered by this benchmark, supporting its evaluation as an alternative to bloXroute; the conclusion remains limited to the stated test periods, regions, network conditions, and measurement rules.

Benchmark Data and Measurement Notes

• This article uses like-for-like BlockRazor and bloXroute tests from Virginia, Frankfurt, Dublin, and Tokyo. Each regional benchmark ran for six hours between September 23 and 25, 2026.
• Node Stream first arrive share assigns each block to the source that reached the test client first, with other P2P sources counted separately. Lead-time percentiles include only samples where BlockRazor arrived first and do not represent end-to-end service latency.
• Block Stream uses the first provider to deliver each observed block as the decision rule. Virginia, Frankfurt, Dublin, and Tokyo contained 47,944, 47,862, 47,981, and 47,668 first arrive decisions, respectively.
• Mempool transaction first arrive compares only transactions observed by both services. Observed share uses the union of unique transaction hashes seen by both providers during the test as its denominator, so it is not a transaction-inclusion rate.
• The public transaction test used paired submissions. The earlier inclusion block determined the winner; if both transactions entered the same block, the lower transaction index won. The test contained 797 valid pairs. Same-block average and maximum index leads are calculated separately from the samples won by each provider.
• All results are specific to the test periods, regions, network conditions, and decision rules described above. They should not be treated as fixed performance differences across every deployment environment.