Spurious 429 'Extra usage is required for long context requests' on Max plan with 1M context

Description

Getting a transient 429 error "Extra usage is required for long context requests" on a Max plan with 1M context enabled. The request was only ~23K tokens, well within limits.

Debug Log Details

• Claude Code version: 2.1.90
• Model: claude-opus-4-6 (1M context)
• Token count at time of error: ~22,905 tokens
• Effective window: 992,000 tokens
• Request ID: req_011CZf8TJf84hAUziB6LuRoc
• Client request ID: a6752fb6-27cc-4a68-a536-7cd61dc995f4

2026-04-02T14:48:58.696Z [DEBUG] autocompact: tokens=22905 threshold=979000 effectiveWindow=992000
2026-04-02T14:48:58.707Z [DEBUG] [API REQUEST] /v1/messages x-client-request-id=a6752fb6-27cc-4a68-a536-7cd61dc995f4 source=repl_main_thread
2026-04-02T14:48:58.867Z [ERROR] API error (attempt 1/11): 429 429 {"type":"error","error":{"type":"rate_limit_error","message":"Extra usage is required for long context requests."},"request_id":"req_011CZf8TJf84hAUziB6LuRoc"}
2026-04-02T14:48:58.872Z [ERROR] Error in API request: 429 {"type":"error","error":{"type":"rate_limit_error","message":"Extra usage is required for long context requests."},"request_id":"req_011CZf8TJf84hAUziB6LuRoc"}

Expected Behavior

The request should succeed — the user has a Max plan with 1M context enabled, and the request was only ~23K tokens.

Actual Behavior

The request fails with a 429 rate limit error claiming extra usage is required for long context. The error was not retried. A subsequent new request (triggered by the next user message) succeeded.

Reproduction

Intermittent — happened once during a normal session. The first API request in the session succeeded; the second (with ~23K tokens) failed.

🤖 Generated with Claude Code