Google: Gemma 4 31B Performance
Try Gemma 4 31B on us. Starter credit covers your first requests, then one key reaches all 13 provider routes with automatic failover.
google/gemma-4-31b-it
Measured performance
Continuously sampled p50/p95 time-to-first-token (TTFT), effective throughput, and success rate for Google: Gemma 4 31B. Effective throughput uses provider-reported output tokens over complete request time. Unsupported route and probe-configuration rows are separated from provider downtime, and no prompt or output content is stored.
| Provider | p50 TTFT | p95 TTFT | Effective throughput | Uptime | Config excluded | Availability samples |
|---|---|---|---|---|---|---|
| tinfoil | 1333 ms | 2724 ms | 79 tok/s n=2 | 100.00% | — | 10 |
| regolo | 1364 ms | 1426 ms | 114 tok/s n=1 | 66.67% | — | 3 |
| siliconflow | 1674 ms | 13725 ms | 40 tok/s n=2 | 100.00% | — | 2 |
| parasail | 1922 ms | 1922 ms | 44 tok/s n=1 | 100.00% | — | 1 |
| sail-research | 2423 ms | 2883 ms | — | 100.00% | — | 3 |
| friendli | 2536 ms | 4265 ms | 364 tok/s n=2 | 100.00% | 1 provider_error |
3 |
| wandb | 4290 ms | 4290 ms | 190 tok/s n=2 | 100.00% | — | 1 |
| crusoe | — | — | 157 tok/s n=1 | — | — | 0 |
| deepinfra | — | — | 59 tok/s n=1 | — | — | 0 |
| greenference | — | — | 90 tok/s n=1 | — | — | 0 |
| io-net | — | — | 91 tok/s n=2 | — | — | 0 |
| novita | — | — | 31 tok/s n=1 | — | — | 0 |
Full provider & model leaderboard.
13 routes.
More routes give the auto router more room to fail over around provider 429 and 5xx responses.
Gateway overhead is measured separately.
Public status separates TLS/health overhead from full model latency so slow LLMs do not inflate the router metric.
Metadata rollups.
Status samples store latency, outcome, provider, model, route, cost, and region metadata only.
View public status or inspect provider routes.