Email me

AI API errors / Anthropic (Claude)

Anthropic (Claude) · HTTP 400 · checked Oct 24, 2026

Anthropic 400 invalid_request_error: what it means and how to fix it

Do not retry: fix it first
400 invalid_request_error: "thinking.type.disabled" is not supported for this model. Use "thinking.type.adaptive" and "output_config.effort" to control thinking behavior.

What it meansOn the newest Claude models thinking is always on, so a request that tries to switch it off is rejected. (Claude Sonnet 5.5 uses between_tools as its lowest setting instead.)

How to fix it

  1. Drop thinking: {"type": "disabled"}; use a low effort level to keep thinking short.
  2. To keep thinking out of the response without turning it off, set display: "omitted" on the thinking config.
  3. On Claude Sonnet 5.5, send thinking: {"type": "between_tools"} for the lowest setting.

Handle it in code

import anthropic

# The SDK retries 429, 5xx and connection errors with exponential backoff and honors retry-after.
client = anthropic.Anthropic(max_retries=6)

try:
    msg = client.messages.create(model="claude-sonnet-5-5", max_tokens=1024,
                                 messages=[{"role": "user", "content": "Hello"}])
except anthropic.RateLimitError as e:      # 429 after all retries
    print("still rate limited", e.response.headers.get("retry-after"))
except anthropic.APIStatusError as e:      # 529 overloaded, 500, 504...
    print(e.status_code, e.response.headers.get("request-id"))

The official SDKs already retry connection errors, 429 and 5xx twice with exponential backoff and honor retry-after; raise max_retries for batch jobs.

Same kind of error at other providers

Other Anthropic (Claude) errors

Source

Paying too much per call while you are here? Compare AI model prices or estimate your bill.

Up next

Paste any errorEvery AI API error, retry or fixAnthropic, OpenAI and Gemini in one finder.Read it next →

Did this answer your question?

Up next

Get 3 picks for your job, volume and budgetPick an AI model for your jobLLM API pricing for 71 models per 1M tokens, checked at the source. Pick for me: your job, volume and budget give 3 models with the monthly cost. Filters, CSV.Read it next →

Get the receipts by email.

What my AI agents tried this week, with the real numbers: what worked, what flopped, what made money. One email when there is something worth reading.

No spam, never sold. Every email has a one-click unsubscribe.

Did this answer your question?

Popular in the shop

All 124 tools →

Browse: Bundles and kitsHome, family and schoolMoney and businessAI toolsSports and fitnessMaking and hobbiesFood and drinkIndie dev and launchTrading and quant

Free: Landing page teardown iPhone app cost calculator LLM API cost calculator Paid online: $49 Launch Check App builds Articles FAQ