Home Overview Platform Architecture Performance & Benchmarks Pricing & ROI Calculator Global Edge Network Enterprise Trust Center Company & US Offices Contact & Support Launch Cloud Project →

Reproducible Benchmark Suite (Q3 2026)

No synthetic marketing claims. We publish our full test scripts, raw telemetry JSONs, and hardware configurations so you can reproduce these results on your own cluster.

$ curl -fsSL https://benchmarks.vectis.cloud/reproduce.sh | bash

Comprehensive Latency Percentiles (500k Concurrent Requests) Ashburn • US-East-1 Core

Provider & Architecture P50 (Median) P90 P99 (Tail) P99.9 (Spike) Error Rate
◆ Vectis Cloud (eBPF + Firecracker) 0.64ms 1.12ms 1.84ms 3.40ms 0.0000%
Cloudflare Workers (V8 Isolate) 2.40ms 5.80ms 8.92ms 19.4ms 0.0012%
Fly.io Machines (Firecracker) 4.80ms 11.2ms 18.50ms 42.1ms 0.0048%
AWS API Gateway + Lambda 8.20ms 16.9ms 24.60ms 88.4ms 0.0120%
Hardware Rig Specification: Dual AMD EPYC 9654 96-Core (192 threads), 768GB DDR5 ECC 4800MHz Memory, 4x 7.68TB Samsung PM9A3 NVMe Gen4 in RAID-0, Dual Mellanox ConnectX-6 Dx 100GbE NICs with RoCEv2 enabled.