Google Vertex
Hosted API- US East (Virginia)3.9 s, view the regional leaderboard
- US West (Oregon)3.9 s, view the regional leaderboard
- EU West (Ireland)3.9 s, view the regional leaderboard
- Asia Pacific (Tokyo)3.9 s, view the regional leaderboard
- Asia Pacific (Mumbai)3.9 s, view the regional leaderboard
- Models on live video
- 2
- Regions measured
- 5
- Real-time models
- 1 of 2
- Median output speed
- 176 tok/s
- Median E2E · US East
- 3.9 s
- Median blended $/1M
- $5.17
Models on Google Vertex
| # | Details | |||||||
|---|---|---|---|---|---|---|---|---|
| 1 | View gemini-3-flashGemini 3 Flash | 247 tok/s | 53 ms | 150 ms | $4.52 | <200 ms | ||
| 2 | View gemini-3-1-proGemini 3.1 ProReasoning | 106 tok/s | 125 ms | 7.6 s | $5.83 | – |
Reasoning models are indicated by a lightbulb icon
Real-time coverage
US East (Virginia)1 of 2 under 200 ms
US West (Oregon)1 of 2 under 200 ms
EU West (Ireland)1 of 2 under 200 ms
Asia Pacific (Tokyo)1 of 2 under 200 ms
Asia Pacific (Mumbai)0 of 2 under 200 ms
Nothing under the bar here yet.
Latency columns use the US East (Virginia) reference region. Real-time marks a model that closes the loop under 200 ms end to end on Google Vertex from at least one region.