Is Claude down?
Live Anthropic API status, pulled from the official status page. If it's green here, the API is up — so the problem is probably on your side.
Is it Claude or is it you?
Before assuming an outage, check the error you're getting — most "Claude not working" reports are client-side:
- 429 Too Many Requests — a rate limit. Your side. Back off and retry with exponential backoff, or request a higher tier.
- 401 / 403 — an invalid or expired API key, or wrong org/project. Your side.
- 400 — a malformed request (bad model string, oversized prompt). Your side — check the model reference.
- 500 / 502 / 503 / timeouts across all calls — usually Anthropic's side. Confirm on the status page.
What to do during a Claude outage
- Fail over. Route to another provider — compare drop-in alternatives by price and capability.
- Retry smart. Exponential backoff with jitter; don't hammer a recovering API.
- Queue non-urgent work and replay it once the status page goes green.
- Subscribe to the Anthropic status page for incident updates.
Anthropic models & pricing
If you're switching or planning a fallback, here's what Anthropic offers.
| Model | Input /1M | Output /1M | Cost / call* | Context | Type |
|---|---|---|---|---|---|
| Claude Haiku 4.5 Fastest, near-frontier | $1.00 | $5.00 | $0.0060 | 200K | Fast / cheap |
| Claude Sonnet 4.6 Best speed/intelligence; caching cuts input ~90% | $3.00 | $15.00 | $0.018 | 1M | Balanced |
| Claude Opus 4.8 Top Opus reasoning/agentic | $5.00 | $25.00 | $0.030 | 1M | Flagship |
| Claude Fable 5 Most capable widely released | $10.00 | $50.00 | $0.060 | 1M | Flagship |
*Reference cost of one call with 1,000 input + 1,000 output tokens — a neutral yardstick. Use the calculator for your real usage. Cheapest row highlighted.
Is Claude down — FAQ
Is Claude down right now?
Check the live status above — it reads Anthropic's official status page (https://status.claude.com). Green means all systems operational; amber or red means a degradation or outage on Anthropic's side, not yours.
How do I know if it's Claude or my own code?
A 429 means you hit a rate limit (your side — back off and retry). 401/403 is an auth or API-key problem (your side). Widespread 500/502/503 errors or timeouts usually mean Anthropic is having an incident — confirm on the status page.
Where is the official Anthropic status page?
https://status.claude.com — it lists current incidents, per-component status (API, console, etc.) and a history of past outages with timelines.
What should I do during a Claude outage?
Fail over to another model or provider if your app supports it (compare alternatives here on AI API Prices), queue non-urgent requests, retry with exponential backoff, and subscribe to the status page for updates.