KEY TAKEAWAYS
- Median latency drives perceived speed; p95 drives error rates on large runs.
- Classified percentage matters more than raw speed.
- Aggregated from 278 real checks over the last 90 days (updated October 14, 2026).
How fast a checker responds matters almost as much as its accuracy: it determines time-to-first-result, how quickly a large list drains and how the tool feels in practice. This page publishes our own field numbers — aggregated from real checks run through this platform over the last 90 days. It updates automatically as new checks arrive.
Benchmarks at a glance
| Gateway | Checks | Median | p95 | Classified |
|---|---|---|---|---|
| Fast Checker | 84 | 8,630 ms | 26,676 ms | 100% |
| Pre-Auth Gateway 2 | 61 | 9,345 ms | 15,964 ms | 100% |
| Gate3 Charged | 42 | 10,203 ms | 18,009 ms | 100% |
| Pre-Auth Gateway | 91 | 13,746 ms | 23,632 ms | 100% |
Median is the typical response time; p95 shows the slow tail that drives timeouts. "Classified" is the share of checks that returned a definitive live/dead answer rather than an error.
How to read these numbers
- Median dominates perceived speed for small lists.
- p95 dominates large runs: if it approaches the gateway timeout, errors climb.
- Classified % is the real reliability metric — a fast gateway that never answers is useless.
Overall median across all gateways in this window: 10,971 ms, p95 22,855 ms. Because these are real-world measurements, they move with provider conditions; treat them as a guide, not a guarantee.
Methodology
Every check records its own latency in milliseconds. This page aggregates per gateway over the last 90 days and excludes gateways with fewer than 10 checks. No customer data is used or published.
Why streaming changes the equation
A batch that takes a minute end-to-end can still deliver its first result in seconds when results are streamed. See streaming results explained and how to choose a gateway.