For creatorsMake videos about AI coding? Earn 50% for life.Earn 50% for life.Become a Morphy creatorJoin →
morphy

Claude Code "API Error: 529 Overloaded": what it means and how to keep working

Claude Code says "API Error: Repeated 529 Overloaded errors" or "Opus is experiencing high load"? What a 529 means, why it isn't your usage limit, and five ways to keep working, from /model to a fallback model.

Short answer: a 529 means Anthropic's servers are at capacity for everyone right now. It isn't your fault, it isn't your usage limit, and it doesn't count against your plan. Claude Code has already retried up to 10 times before it shows you the error. To keep working, switch to another model with /model (capacity is tracked per model), wait a few minutes and type try again, or set a fallback model so Claude Code switches by itself next time.

Everything below is from Anthropic's own documentation, checked on October 28, 2026. (I build Morphy, which runs Claude Code next to other AIs; it comes up once, near the end.)

The messages

In Claude Code, after its retries run out:

API Error: Repeated 529 Overloaded errors. The API is at capacity — this is usually temporary. Try again in a moment. If it persists, check https://status.claude.com.

Some versions word it like this:

API Error: 529 Overloaded. This is a server-side issue, usually temporary — try again in a moment. If it persists, check https://status.claude.com.

When one model is under especially high load, Claude Code tells you which one and what to switch to:

Opus is experiencing high load, please use /model to switch to Sonnet

In the Claude desktop app's Code tab it reads Opus is experiencing high load. Switch to Sonnet. On Fable models, the message names Fable.

If you call the API yourself, a 529 comes back as an error of type overloaded_error: "The API is temporarily overloaded."

What a 529 means

The API is temporarily at capacity across all users. In Anthropic's words: "A 529 is not your usage limit and doesn't count against your quota." Nothing in your prompt, settings or account caused it.

Two things follow from how it works:

  • It's per model. Capacity is tracked per model, so when Opus is overloaded, Sonnet often isn't.
  • Claude Code already tried. It retries failures like this up to 10 times with exponential backoff before you see anything, so by the time the error shows, an immediate retry won't do much.

Five ways to keep working

  1. Check the status page. status.claude.com shows capacity notices and incidents. If you use Claude through Amazon Bedrock, Google Cloud or a gateway, the message names that provider's status page instead.

  2. Switch models. Run /model and pick another one, usually Sonnet when Opus is overloaded. In the desktop app, use the model picker. Switch back later with /model again.

  3. Wait a few minutes, then send try again. Your message is still in the conversation, so you don't need to paste a long prompt again.

  4. Set a fallback model, so it switches by itself. When the main model is overloaded, Claude Code tries the next one in your list, for that turn only:

    claude --fallback-model sonnet,haiku

    Or keep it in your settings for every session:

    {
      "fallbackModel": ["claude-sonnet-5", "claude-haiku-4-5"]
    }

    Up to three models; your next message tries the main model first again. Claude Code doesn't show the chain at startup or in /status: the notice when it switches is the first sign it's set.

  5. For scripts and CI, wait it out. Set CLAUDE_CODE_RETRY_WATCHDOG=1 and Claude Code keeps retrying 429 and 529 capacity errors instead of failing after its usual attempts. To spread the normal retries over a longer window instead, raise CLAUDE_CODE_OVERLOADED_RETRY_BASE_DELAY_MS (default 500, up to 32,000) and CLAUDE_CODE_OVERLOADED_RETRY_MAX_DELAY_MS (default 32,000, up to 64,000).

And if it keeps happening on a busy day, keep a second agent on hand: Codex, on a ChatGPT plan, has its own limits and its own servers.

529, 429, 500 or your usage limit?

They look alike but mean different things:

MessageWhat it meansWhat to do
Repeated 529 Overloaded errorsAnthropic is at capacity for everyoneSwitch models, or wait a few minutes
Server is temporarily limiting requests (not your usage limit)A short-lived throttle on Anthropic's sideWait briefly and try again
Request rejected (429)The rate limit on your API key or cloud projectCheck /status for the credential in use; lower how many agents run at once
500 Internal server errorAn unexpected failure inside the APIWait a minute and try again
You've hit your session limit · resets 3:45pmYour plan's own usage limitWait for the reset, or use another plan; see Claude Code's usage limits

One trap with Request rejected (429): a stray ANTHROPIC_API_KEY in your environment can send requests through a low-tier API key instead of your subscription. /status shows which credential is active.

If you call the API yourself

Anthropic's SDKs retry 529s and other temporary errors twice by default, with backoff; raise max_retries (maxRetries in TypeScript) to retry more. A related trap: if your organization's usage jumps sharply, the API can answer with 429 errors from its acceleration limits. Anthropic's advice is to increase traffic gradually and keep usage steady.

In Morphy

When a Claude chat stops on a 529 in Morphy, you can switch that chat to another AI without starting over: open the model button, point at Provider and pick Codex (or any AI you've connected). Your next message brings the conversation along, and when Claude is back, switching back hands Claude only what happened since. See Switch to another AI.

Questions

Does a 529 count against my Claude usage limit? No. Anthropic says a 529 "is not your usage limit and doesn't count against your quota."

Is the 529 error my fault? No. It means the API is at capacity across all users; it isn't caused by your prompt, settings or account.

How long does a 529 last? Anthropic calls it "usually temporary". Check status.claude.com for a current incident, and switch models in the meantime.

Does switching to Sonnet help? Often, yes. Capacity is tracked per model, and Claude Code itself suggests Sonnet when Opus is under high load.

What's the difference between a 529 and a 429? A 529 is Anthropic's servers at capacity for everyone. A 429 is a rate limit: on your API key or cloud project, or a short throttle that Claude Code labels "not your usage limit".

References