About LatencyRadar

LatencyRadar runs independent measurements from locations across Europe, North America, Asia, and other regions.

For each request, we can measure the individual parts that contribute to latency, including:

• DNS lookup
• Connection time
• TLS handshake
• Time to first byte (TTFB)
• Total response time

For AI APIs, we also track metrics such as time to first token where applicable.

By collecting these measurements continuously, LatencyRadar can show not only how a service performs today, but how that performance changes over time and across regions.

A core part of LatencyRadar is its public benchmark dataset.

We regularly measure APIs and services under consistent conditions so developers can answer questions such as:

• Which AI providers respond fastest?
• Does a provider perform differently in Europe, Asia, or the Americas?
• Has an API become slower over the last week or month?
• How much latency comes from the network versus the provider itself?
• How does performance change when using a gateway or intermediary?

The aim is not to rank companies based on marketing claims, but to publish observable measurements and the methodology behind them.

Commercial relationships, sponsorships, or paid research do not change benchmark results or conclusions.

LatencyRadar also lets developers measure their own websites and APIs from the same global infrastructure used for public benchmarks.

This makes it possible to compare your own performance with the wider ecosystem and identify regional problems before they become user complaints.

No account is required to run a basic test.

A single latency test is useful.

A continuously collected history is much more useful.

LatencyRadar is building a historical dataset of API performance across providers, regions, and time. As that dataset grows, it becomes possible to identify regressions, regional differences, infrastructure changes, and longer-term trends that are difficult to see from isolated tests.

The long-term vision is:

LatencyRadar measures how the APIs the internet depends on perform around the world, and lets you monitor your own alongside them.