Is OpenRouter Down?
Live OpenRouter status. Auto-refreshes every 2 minutes.
OpenRouter is Operational
OpenRouter is up and running normally. All systems are operational.
Official status page: status.openrouter.ai
Component Status
Recent OpenRouter incidents
What to do when OpenRouter is down
Call the upstream provider directly
OpenRouter sits in front of the real model owners, so when the routing layer is down the providers themselves are usually fine. Point your client at the native API for the model you need (Claude, GPT, Gemini, DeepSeek, and friends), since most accept the same OpenAI-style requests you already send. Check each provider's live status first so you switch to one that is actually up.
Switch to a sibling inference gateway
If you want to keep the convenience of one endpoint and one key, other aggregators host overlapping catalogs on completely separate infrastructure. Together and Fireworks both serve large open-weight and frontier-adjacent model lineups, and Groq is a fast option for the open models it carries. Compare live status and pick whichever is green right now.
Get notified when status changes
Instead of refreshing this page, subscribe to TensorFeed alerts and get pinged the moment OpenRouter (or any upstream provider it routes to) flips between operational, degraded, and down. Pair it with the live status board and the provider leaderboard to pick the most reliable route at any moment.
Frequently Asked Questions
Is OpenRouter down right now?
The live indicator at the top of this page shows the current state. A green dot means the OpenRouter routing API is operational, amber means degraded (requests may be slow or some upstream providers are unreachable), and red means a confirmed outage. The signal is pulled from OpenRouter's official status page and refreshes every couple of minutes, so read the dot and the last-checked timestamp rather than guessing from a single failed request.
How do I check OpenRouter status?
Watch the live indicator on this page, which mirrors OpenRouter's official status page at status.openrouter.ai. That page breaks out uptime for the chat completions API, the models and data APIs, and the website separately, so you can tell whether the routing layer itself is failing or just one component. If your own requests fail but every indicator is green, the problem is more likely your API key, credit balance, or a single upstream provider.
What should I do when OpenRouter is down?
Because OpenRouter is only a routing layer, the upstream providers behind it are usually still online when OpenRouter itself fails. The fastest fix is to point your client directly at the model owner's native API (Anthropic, OpenAI, Google, DeepSeek, and others) since most are compatible with the same OpenAI-style request format you already use. If you need to keep aggregator convenience, sibling gateways like Together or Fireworks host overlapping catalogs on independent infrastructure.
How often does OpenRouter go down?
OpenRouter is generally reliable, and because it can fail over across multiple upstream providers, a single provider outage often does not surface as downtime for you at all. The incidents that do affect everyone tend to be problems with OpenRouter's own routing, auth, or billing layer rather than the models. Brief degraded windows (slower routing, one or two unreachable upstreams) are more common than full outages. Check the official status history for the real track record rather than relying on anecdotes.
Which models and providers does OpenRouter cover, and what is the common confusion?
OpenRouter aggregates hundreds of models, including Anthropic Claude, OpenAI GPT, Google Gemini, Meta Llama, DeepSeek, Mistral, Qwen, and many open-weight options, all behind one OpenAI-compatible endpoint and one API key. The most common confusion is blaming OpenRouter for an error that actually originates upstream: if a specific model fails while everything else works, the issue is usually that model's underlying provider, not OpenRouter's routing. OpenRouter also exposes a free model tier and per-model pricing, so some failures are simply rate limits or credit issues, not outages.
Where can I see OpenRouter incident history?
OpenRouter's official status page at status.openrouter.ai keeps a running log of past incidents with timestamps, affected components, and resolution notes. For a broader view across the whole stack, TensorFeed tracks OpenRouter alongside every major upstream provider so you can correlate an OpenRouter blip with the model vendor that likely caused it. Subscribe to alerts to get notified the moment the status changes instead of refreshing manually.