Meta: Llama 3.3 70B Instruct Performance
Try Llama 3.3 70B Instruct on us. Starter credit covers your first requests, then one key reaches all 10 provider routes with automatic failover.
meta-llama/llama-3.3-70b-instruct
Measured performance
Continuously sampled p50/p95 time-to-first-token (TTFT), effective throughput, and success rate for Meta: Llama 3.3 70B Instruct. 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 |
|---|---|---|---|---|---|---|
| telnyx | 889 ms | 889 ms | — | 100.00% | — | 1 |
| databricks | 1052 ms | 1870 ms | — | 100.00% | — | 4 |
| akashml | 1082 ms | 1093 ms | — | 100.00% | — | 5 |
| tinfoil | 1098 ms | 1345 ms | — | 100.00% | — | 3 |
| parasail | 1190 ms | 1190 ms | — | 100.00% | — | 1 |
| novita | 1258 ms | 1542 ms | — | 100.00% | — | 2 |
| io-net | 1332 ms | 1332 ms | — | 100.00% | — | 1 |
| wandb | 1401 ms | 1401 ms | — | 100.00% | — | 1 |
| scaleway | 1484 ms | 2001 ms | — | 100.00% | — | 3 |
| sambanova | — | — | — | — | 7 provider_error |
0 |
Full provider & model leaderboard.
10 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.