- Score
- 99.5
- p95
- 46 ms
Benchmarks
Which one should I use
Five questions, and a shortlist filtered from what has actually been measured. This page invents no score: it rules out providers that cannot do what you asked, and orders the rest by the same composite every other page shows — with the measurement behind every position, and behind every exclusion.
Use one of these
ordered by the published score · 24h window- Score
- 97.4
- p95
- 54 ms
- Score
- 83.4
- p95
- 100 ms
- Score
- 78.5
- p95
- 122 ms
- Score
- 77.4
- p95
- 128 ms
- Score
- 77.4
- p95
- 128 ms
- Score
- 69.4
- p95
- 202 ms
- Score
- 49.5
- p95
- 93 ms
- Score
- 31.9
- p95
- 5243 ms
What this page does and does not do
It invents no score
Your answers do not produce a bespoke ranking. That would be a second scoring system sitting beside the published one. This filters and orders; the order is the same one the chain page shows.
Unmeasured is not unsupported
A provider that has not been asked about a method has not failed it. The same goes for a transport we hold no endpoint of theirs on. Ruling either out would let measurement coverage decide your shortlist, so they sit in their own list and say so.
Every exclusion carries its evidence
A filter you cannot argue with is a filter you cannot trust. Each ruled-out provider names the measurement that ruled it out, and links to the page it came from.
Price is not measured, so it is not ranked
“Free” here means a public endpoint that needs no key — a fact about the endpoint, observable from here. What a paid plan costs is not, and this platform does not publish figures it cannot check. How everything here is measured