Some services degraded

Uptime
100.00%
90 days
First token
2252ms
Median
Healthy routes
343/357
Open incidents
0
Models

Model availability

Live routes per model family.

45 families

Current model family health. Families needing attention appear first. Availability is the share of routes currently serving, including degraded routes.
Model familyStatusRoutesAvailabilityHealthyDegradedOpen circuitServing networks
AbliteratedDegraded66.67%2/301abliteration
alibabaDegraded100.00%31/3210alibaba, darkbloom, vercel
anthropicDegraded100.00%18/1910vercel
Bonsai 2Degraded100.00%0/110darkbloom
deepseekDegraded100.00%13/1410vercel
googleDegraded100.00%22/2310darkbloom, vercel, vertex
inclusionaiDegraded100.00%5/610vercel
moonshotaiDegraded100.00%8/910vercel
openaiDegraded100.00%74/7730darkbloom, openai, vercel
stepfunDegraded100.00%1/210vercel
xiaomiDegraded100.00%2/310vercel
zaiDegraded100.00%17/1810vercel
amazonOperational100.00%4/400vercel
bflOperational100.00%4/400bfl
bytedanceOperational100.00%8/800bytedance, vercel
callstackOperational100.00%1/100vercel
cohereOperational100.00%1/100vercel
elevenlabsOperational100.00%8/800vercel
elorianOperational100.00%1/100vercel
fireworksOperational100.00%1/100vercel
fish-audioOperational100.00%5/500vercel
inceptionOperational100.00%1/100vercel
inference-netOperational100.00%2/200vercel
interfazeOperational100.00%1/100vercel
klingaiOperational100.00%6/600klingai
liquidOperational100.00%3/300vercel
meituanOperational100.00%1/100vercel
metaOperational100.00%15/1500meta, vercel
microsoftOperational100.00%6/600vercel
minimaxOperational100.00%11/1100minimax, vercel
mistralOperational100.00%12/1200vercel
mixedbreadOperational100.00%1/100vercel
morphOperational100.00%2/200vercel
nvidiaOperational100.00%7/700vercel
paperinstrumentsOperational100.00%1/100vercel
particleOperational100.00%1/100vercel
perplexityOperational100.00%3/300vercel
Qwen3-VLOperational100.00%1/100darkbloom
Qwen3.5Operational100.00%1/100darkbloom
recraftOperational100.00%9/900recraft
sakanaOperational100.00%4/400vercel
spacexaiOperational100.00%17/1700vercel, xai
tencentOperational100.00%6/600vercel
thinkingmachinesOperational100.00%3/300vercel
typesafe-aiOperational100.00%3/300vercel

Availability counts healthy and degraded serving routes, not historical uptime. Degraded routes can trigger fallback.

Providers

Provider checks

One generation check per provider, every five minutes.

ProviderStatusTest modelLatencyLast checked
inceptionDegradedinception/mercury-3325ms
sakanaDegradedsakana/namazu55ms
alibabaOperationalalibaba/qwen3-coder-30b-a3b362ms
amazonOperationalamazon/nova-micro976ms
anthropicOperationalanthropic/claude-haiku-5.5593ms
bytedanceOperationalbytedance/seed-2.1-turbo3627ms
callstackOperationalcallstack/apex847ms
cohereOperationalcohere/command-a373ms
darkbloomOperationalalibaba/qwen3.5-9b2116ms
deepseekOperationaldeepseek/deepseek-v4-flash1494ms
elorianOperationalelorian/glimpse1745ms
fireworksOperationalfireworks/ember-1874ms
googleOperationalgoogle/gemini-2.5-flash-lite490ms
inclusionaiOperationalinclusionai/ling-3.0-flash1025ms
inference-netOperationalinference-net/schematron-v2-turbo728ms
interfazeOperationalinterfaze/interfaze-beta1970ms
liquidOperationalliquid/lfm2.5-24b-a2b631ms
meituanOperationalmeituan/longcat-2.52310ms
metaOperationalmeta/llama-3.1-8b344ms
minimaxOperationalminimax/minimax-m3.1-flash931ms
mistralOperationalmistral/mistral-nemo1350ms
mixedbreadOperationalmixedbread/toast-1892ms
moonshotaiOperationalmoonshotai/kimi-k2792ms
nvidiaOperationalnvidia/nemotron-3.5-lightning-free368ms
openaiOperationalopenai/gpt-4.1-nano656ms
paperinstrumentsOperationalpaperinstruments/ultramarine-1-flash1119ms
particleOperationalparticle/particle-1690ms
perplexityOperationalperplexity/sonar1763ms
spacexaiOperationalspacexai/grok-4.20-non-reasoning592ms
stepfunOperationalstepfun/step-3.7-flash2122ms
tencentOperationaltencent/hy-mt2-lite1395ms
thinkingmachinesOperationalthinkingmachines/inkling-2-small663ms
xiaomiOperationalxiaomi/mimo-v2.6-flash2456ms
zaiOperationalzai/glm-4.7-flashx1167ms

A passing check does not guarantee every model is available.

Reliability

Measured reliability

Gateway reliability

WindowError rateRequests
24 hours0.00%16,579
7 days0.02%159,402
90 days0.08%1,294,921

First token: 2252ms median · 23647ms p95

Upstream failures, timeouts and exhausted retries. Excludes cancellations and refusals. No requests means no error rate.

Deposit crediting

MeasurementDelay
Median lag0s
95th percentile0s
Oldest uncredited—

From on-chain finality to credit in your balance.

Incidents

Incident history

No incidents recorded.