OpenAI compatible API · Attested · Public status

DeepSeek: DeepSeek V4 Flash 0731 Performance

Compare measured TTFT, TTFB, throughput, uptime, and route health for DeepSeek V4 Flash 0731 across TrustedRouter providers using metadata-only production probes.

Verify gateway
Onebase URL to migrate
100sof models and routes
0prompt or output logs. Always.

deepseek/deepseek-v4-flash-0731

open weights Performance

All models

Measured performance

Continuously sampled p50/p95 time-to-first-token (TTFT), time-to-first-byte (TTFB), effective throughput, and success rate for DeepSeek: DeepSeek V4 Flash 0731. 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.

Providerp50 TTFTp95 TTFTp50 TTFBEffective throughputUptimeConfig excludedAvailability samples
deepseek 1204 ms 3702 ms 1738 ms 132 tok/s n=2 100.00% 166
novita 1222 ms 1486 ms 1222 ms 86 tok/s n=1 100.00% 2
siliconflow 1533 ms 1710 ms 1533 ms 57 tok/s n=1 100.00% 3
baseten 1767 ms 2630 ms 1766 ms 177 tok/s n=2 100.00% 5
deepinfra 2659 ms 13849 ms 2659 ms 53 tok/s n=1 100.00% 2
fireworks 2891 ms 23396 ms 2891 ms 44 tok/s n=2 100.00% 9
atlas-cloud 47 tok/s n=1 0
together 74 tok/s n=1 4 probe_config_error 0

Full provider & model leaderboard.

Provider diversity

16 routes.

More routes give the auto router more room to fail over around provider 429 and 5xx responses.

Streaming

Gateway overhead is measured separately.

Public status separates TLS/health overhead from full model latency so slow LLMs do not inflate the router metric.

Status

Metadata rollups.

Status samples store latency, outcome, provider, model, route, cost, and region metadata only.

Workspace access

Sign in

Choose a sign in method to access your TrustedRouter workspace.

By signing in you agree to the terms of service and privacy policy.