Anthropic 529 overloaded_error

Why am I getting Anthropic 529 overloaded_error?

A Messages API call returns HTTP 529 with an `overloaded_error` body — not a rate-limit response tied to your key, but a capacity signal from Anthropic itself, and it can happen on a request well inside your own tier limits.

{"type":"error","error":{"type":"overloaded_error","message":"Overloaded"}}

529 is Anthropic's own capacity backpressure, distinct from the per-key 429 rate limit — it fires when Anthropic's infrastructure is under load platform-wide, independent of your usage tier or how many requests you've sent. There is no equivalent status code on OpenAI's API, which returns a 429 for every kind of throttling instead of separating the two causes.

async function callWithBackoff(maxRetries = 5) {
  for (let attempt = 0; attempt < maxRetries; attempt++) {
    try {
      return await client.messages.create({
        model: MODEL,
        max_tokens: 1024,
        messages: [{ role: 'user', content: 'Say hello in one sentence.' }],
      });
    } catch (err) {
      if (err instanceof Anthropic.APIError && err.status === 429) {
        const retryAfter = Number(err.headers?.['retry-after'] ?? 2 ** attempt);
        await new Promise((r) => setTimeout(r, retryAfter * 1000));
        continue;
      }
      throw err;
    }
  }
  throw new Error('exceeded max retries');
}

const response = await callWithBackoff();
console.log(response);

Don't retry immediately in a tight loop — a burst of retries against a capacity or rate signal just adds load and delays recovery. Use jittered exponential backoff.

Reproduced against Anthropic's own documentation and the fixture at code-examples/javascript/anthropic/errors.ts on 2026-08-08; not yet observed in production traffic.