Is Groq Down?
Live Groq API status from GroqCloud. Auto-refreshes every 2 minutes.
Groq is Operational
Groq is up and running normally. All systems are operational.
Component Status
Recent Groq incidents
What to do when Groq is down
Groq serves open-weight models, so failover is unusually clean: the same model often runs on another provider. Three practical options while you wait, ranked by closeness of the fallback for typical Groq use cases:
For speed: Cerebras, Together, or Fireworks
All three serve the same open weights at high throughput. Compare them on our inference providers page. Live status: Is Together down? and Is Fireworks down?
For automatic failover: route through OpenRouter
OpenRouter falls back across providers when any single one returns errors. See the OpenRouter status and the model catalog for the full list of routable models.
Keep the model, swap the base URL
Because Groq runs open weights, you usually do not need to change your prompts or model choice during an outage, only the provider endpoint. Browse the same models on open-weights models to confirm which providers carry your model.
Get notified when status changes
Subscribe to TensorFeed outage alerts to get an email the moment Groq (or any tracked AI service) goes degraded or down, and again when it recovers. Free, no account required.
Frequently Asked Questions
Is Groq down right now?
No, Groq is not down right now. GroqCloud and the API are operational and serving requests normally.
How do I check if Groq is down?
Three options: (1) this page, which polls Groq every two minutes, (2) groqstatus.com directly, (3) the @GroqInc account on X for incident commentary. Our page combines the official status feed with per-model component detail and adds historical context the official page does not show.
Is Groq the same as Grok?
No, and the names are confused constantly. Groq (with a q) is an inference hardware company whose LPU chips serve open models like Llama, Kimi, and Gemma at very high token speeds through GroqCloud. Grok (with a k) is xAI's chatbot and model family. Different companies, different products. This page tracks Groq the inference provider.
What do I do when Groq is down?
Groq runs open models, so the cleanest fallback is another inference provider serving the same weights: Together, Fireworks, or Cerebras for speed, or OpenRouter to route around a single provider outage automatically. Point your client at the substitute base URL, keep the model name, and you are usually back in minutes. Subscribe to TensorFeed outage alerts to know the moment status changes.
How often does Groq go down?
Groq has historically maintained high availability across GroqCloud, with most issues being brief per-model degradation (a model pulled for maintenance or elevated latency) rather than full outages. Major platform-wide outages are uncommon; most incidents resolve within 15 to 60 minutes.
Which models does Groq run?
Groq serves open-weight models on its own LPU inference hardware, including Llama 4 variants, Kimi K2, Gemma, and guard models, plus its own compound models. Because these are open weights, the same models are available on other providers, which is why failover is straightforward when Groq has an incident.
Where can I see Groq incident history?
Groq publishes status at groqstatus.com. Our incidents feed at tensorfeed.ai/incidents aggregates incidents across every tracked AI provider in one place, so you can compare Groq reliability against other inference providers like Together, Fireworks, and OpenRouter.