# together vs xai — AI API latency comparison
Updated July 24, 2026. As of July 24, 2026, together has lower edge latency (time-to-first-byte) than xai in 3 of 4 measured region(s). From US (Central), together responds in 118 ms p50 vs 128 ms.

| Region | together p50 | xai p50 | Faster |
| --- | --- | --- | --- |
| Asia (Tokyo) | 186 ms | 172 ms | xai |
| Europe (Germany) | 298 ms | 299 ms | together |
| South America (São Paulo) | 255 ms | 265 ms | together |
| US (Central) | 118 ms | 128 ms | together |


Source: https://llmlatency.dev/compare/together-vs-xai