OpenAI compatible API. Attested gateway. Public status.

Parasail performance

Measured TTFT, TTFB, throughput, uptime, and sampled model routes for Parasail.

Verify gateway
1 URLbase_url migration
100smodels and routes
0prompt logs by default

parasail

272 samples

Provider overview

Continuously sampled provider performance. TrustedRouter reports unsupported route and probe-configuration rows separately from provider downtime. Prompt and output content is not stored.

p50 TTFT992 ms
p95 TTFT ms
p50 TTFB ms
Throughput
Uptime77.94%

Measured model routes

Modelp50 TTFTp50 TTFBThroughputUptimeConfig excludedSamples
thedrummer/cydonia-24b-v4.1 581 ms 578 ms 100.00% 10
qwen/qwen3-next-80b-a3b-instruct 684 ms 682 ms 100.00% 9
thedrummer/skyfall-36b-v2 715 ms 714 ms 100.00% 13
qwen/qwen3-coder-next 784 ms 721 ms 92.31% 13
meta-llama/llama-4-maverick 834 ms 747 ms 100.00% 14
mistralai/mistral-small-3.2-24b-instruct 842 ms 738 ms 100.00% 9
google/gemma-3-27b-it 900 ms 897 ms 100.00% 12
google/gemma-4-26b-a4b-it 948 ms 871 ms 100.00% 4
qwen/qwen3-vl-8b-instruct 954 ms 851 ms 75.00% 4
openai/gpt-oss-20b 956 ms 880 ms 100.00% 18
qwen/qwen3-vl-235b-a22b-instruct 992 ms 972 ms 90.91% 11
bytedance/ui-tars-1.5-7b 992 ms 890 ms 37.50% 8
z-ai/glm-5.1 1002 ms 900 ms 100.00% 7
minimax/minimax-m2.5 1010 ms 931 ms 100.00% 15
openai/gpt-oss-120b 1061 ms 957 ms 100.00% 17
meta-llama/llama-3.3-70b-instruct 1062 ms 957 ms 100.00% 9
qwen/qwen2.5-vl-72b-instruct 1065 ms 981 ms 100.00% 9
google/gemma-4-31b-it 1095 ms 990 ms 100.00% 7
moonshotai/kimi-k2.6 1346 ms 1244 ms 100.00% 14
arcee-ai/trinity-large-thinking 1481 ms 1422 ms 100.00% 17
deepseek/deepseek-v3.2 0.00% 12
z-ai/glm-4.7 0.00% 12
stepfun/step-3.5-flash 0.00% 13
moonshotai/kimi-k2.5 0.00% 8
deepseek/deepseek-v4-pro 0.00% 6 probe_config_error 7

Sign in

Choose a sign in method.