Free tool · China & network
City-to-city internet latency
Median round-trip times between major cities over the last 24 hours, measured from independent public probes. This is public internet latency, not our network latency.
Measured by independent public probes (Globalping). This is public internet latency, not our network.
Each destination target is a public host in that city, not ours.
Measurements start with the next scheduled run. The first figures appear within four hours.
The table will show the median round-trip time for each city pair, the seven-day trend and the fiber floor. A pair appears once at least four measurements exist in the last 24 hours.
A city pair appears once at least four measurements exist in the last 24 hours.
As of Nov 10, 2026, 3:01 AM
| Origin | Beijing | Shanghai | Chengdu | Shenzhen | Hong Kong | Taipei | Singapore | Tokyo |
|---|---|---|---|---|---|---|---|---|
| Toronto | 224.9 ms | 312.4 ms | 252.5 ms | 263.3 ms | 214.6 ms | 188.9 ms | 259.8 ms | 181.6 ms |
| Mexico City | 245.6 ms | 313.2 ms | 335.2 ms | 285.4 ms | 199.5 ms | 178.4 ms | 214.4 ms | 164.5 ms |
| São Paulo | 342.5 ms | 412.5 ms | 403.2 ms | 414.1 ms | 328.7 ms | 299.5 ms | 354.6 ms | 264.3 ms |
| Buenos Aires | 407.8 ms | 454 ms | 406.9 ms | 413.4 ms | 358.5 ms | 318.5 ms | 375.7 ms | 297.1 ms |
| Santiago | 390 ms | 414.8 ms | 372.7 ms | 392.8 ms | 313.8 ms | 296 ms | 371.8 ms | 267.8 ms |
| Lima | 335.4 ms | 375.7 ms | 349.6 ms | 349.4 ms | 286 ms | 253.1 ms | 334.9 ms | 237.6 ms |
| Bogotá | 331.9 ms | 378.8 ms | 376.7 ms | 334.9 ms | 272.9 ms | 233.8 ms | 300.6 ms | 230 ms |
| London | 266.5 ms | 394.2 ms | 318.6 ms | 298 ms | 243.6 ms | 261.6 ms | 248.3 ms | 249.6 ms |
| Frankfurt | 215.7 ms | 351 ms | 266.3 ms | 265.7 ms | 165.3 ms | 277.9 ms | 179.1 ms | 238.4 ms |
| Paris | 221.9 ms | 330.7 ms | 264.3 ms | 262.4 ms | 231.5 ms | 274.1 ms | 179 ms | 251.1 ms |
| Madrid | 256.4 ms | 386.7 ms | 296.4 ms | 272.5 ms | 206.8 ms | 295.4 ms | 265.9 ms | 255.8 ms |
| Amsterdam | 238.8 ms | 330.6 ms | 272 ms | 263.9 ms | 182.7 ms | 279.8 ms | 182.1 ms | 234.3 ms |
| Dubai | 271.5 ms | 252.9 ms | 177.5 ms | 259.3 ms | 200.4 ms | 126.8 ms | 159.8 ms | 222.9 ms |
| Riyadh | 280.2 ms | 427.1 ms | 325.6 ms | 315.4 ms | 239.4 ms | 338.6 ms | 244.8 ms | 303.6 ms |
| Istanbul | 242.3 ms | 389.9 ms | 293.3 ms | 298 ms | 231.7 ms | 313.3 ms | 217.3 ms | 281.5 ms |
| Singapore | 72 ms | 275.4 ms | 103.4 ms | 107.5 ms | 42.8 ms | 49.3 ms | — | 73.1 ms |
| Jakarta | 93.3 ms | 112.6 ms | 118.3 ms | 118.4 ms | 59.4 ms | 133.3 ms | 18 ms | 88.2 ms |
| Bangkok | 228 ms | 131.5 ms | 147.9 ms | 142.4 ms | 67.3 ms | 83.8 ms | 41.5 ms | 109.9 ms |
| Ho Chi Minh City | 252.9 ms | 454.6 ms | 291 ms | 294.6 ms | 36.2 ms | 84.9 ms | 30.7 ms | 110.7 ms |
| Kuala Lumpur | 82.7 ms | 105.5 ms | 106.4 ms | 115.5 ms | 118.2 ms | 209.6 ms | 10.6 ms | 74.2 ms |
Measured by independent public probes (Globalping). This is public internet latency, not our network.
| Toronto | ||||||||
|---|---|---|---|---|---|---|---|---|
| Mexico City | ||||||||
| São Paulo | ||||||||
| Buenos Aires | ||||||||
| Santiago | ||||||||
| Lima | ||||||||
| Bogotá | ||||||||
| London | ||||||||
| Frankfurt | ||||||||
| Paris | ||||||||
| Madrid | ||||||||
| Amsterdam | ||||||||
| Dubai | ||||||||
| Riyadh | ||||||||
| Istanbul | ||||||||
| Singapore | — | |||||||
| Jakarta | ||||||||
| Bangkok | ||||||||
| Ho Chi Minh City | ||||||||
| Kuala Lumpur |
How this works and its limits
Each figure is the median round-trip time over the last 24 hours from independent Globalping probes in the origin city to a public host in the destination city. The target is a university or research-network host, not ours. The table uses 20 measurement origins in 20 countries across the Americas, Europe, the Middle East and Southeast Asia. Origins are measurement points, not a service map. We keep one probe per city pair per hour, spread across the hour.
The table shows the current median, a seven-day sparkline and a change against the previous week. Select a cell for the time series, the probe network, the loss and the fiber floor. A pair with fewer than four samples in the last 24 hours shows no figure rather than an estimate.
The fiber floor is the lower bound for the round trip: twice the great-circle distance times 1.25, divided by 204 km/ms, the speed of light in fiber. Real paths are longer. A measured value below the floor is a data problem, not a fast route.
Data sources and licences
Operator names, brands and codes shown here are public reference data compiled from national regulators and other public sources. Their appearance does not indicate any partnership, endorsement, agreement or connectivity between unicomam.com and the named operators. All trademarks belong to their owners. Data is provided “as is” with no warranty; verify against the official source before relying on it for routing, billing or compliance decisions.
Regulatory information is for reference only and is not legal advice; check the regulator link and the “last verified” date.
Measurements powered by Globalping (opens in a new tab), a public measurement network; results and targets are public under Globalping’s terms.
What changes public city-to-city latency
Distance sets the floor. Light in fiber covers about 204 km per millisecond, and a real cable route is about 1.25 times the straight line. The table shows that floor next to every measured value so the two can be compared.
On top of distance, a packet crosses public networks. Their peering, congestion and the time of day move the median. A seven-day trend says more than one reading. A probe or network change is marked separately, because a different probe is a step change, not a quality change.
For a private path, find a submarine-cable route or ask for a quote. The table itself measures public internet latency only.