Z.ai: GLM 5.3 Performance
Compare measured TTFT, throughput, uptime, and route health for Z.ai: GLM 5.3 across TrustedRouter providers using metadata-only production probes.
z-ai/glm-5.3
Measured performance
Continuously sampled p50/p95 time-to-first-token (TTFT), effective throughput, and success rate for Z.ai: GLM 5.3. 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 |
|---|---|---|---|---|---|---|
| siliconflow | 1126 ms | 1126 ms | — | 100.00% | — | 1 |
| baseten | 1321 ms | 5118 ms | — | 100.00% | — | 2 |
| wafer | 1412 ms | 3560 ms | — | 66.67% | — | 6 |
| makora | 1519 ms | 3174 ms | — | 80.00% | — | 5 |
| akashml | 1548 ms | 1548 ms | — | 100.00% | — | 1 |
| engy | 1813 ms | 2262 ms | — | 100.00% | — | 5 |
| thinkingmachines | 2117 ms | 3919 ms | — | 100.00% | — | 5 |
| atlas-cloud | 2650 ms | 2650 ms | — | 100.00% | — | 1 |
| pearl | 2653 ms | 3248 ms | — | 100.00% | — | 5 |
| fireworks | 2841 ms | 3028 ms | — | 100.00% | — | 5 |
| tinfoil | 2886 ms | 19436 ms | — | 100.00% | — | 5 |
| friendli | 3116 ms | 14091 ms | — | 100.00% | — | 5 |
| zero-g | 3260 ms | 6206 ms | 93 tok/s n=2 | 100.00% | — | 8 |
| sail-research | 3377 ms | 4451 ms | — | 100.00% | — | 3 |
| io-net | 3523 ms | 3523 ms | — | 100.00% | — | 1 |
| deepinfra | 3688 ms | 11260 ms | — | 100.00% | — | 2 |
| inceptron | 3807 ms | 18512 ms | — | 100.00% | — | 7 |
| arcee | 4143 ms | 5600 ms | — | 100.00% | — | 3 |
| zai | 5527 ms | 5527 ms | — | 100.00% | — | 2 |
| nebius | — | — | — | 0.00% | — | 1 |
| phala | — | — | — | 0.00% | — | 7 |
Full provider & model leaderboard.
26 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.