Is this an expected behavior? Sonnet counts both input and max output tokens toward the 200k context window.

[lavantien]

Issue

40k tokens prompt

litellm.BadRequestError: AnthropicException - b'{"type":"error","error":{"type":"invalid_request_error","message":"input length and `max_tokens` exceed context limit: 182831 + 64000 >
200000, decrease input length or `max_tokens` and try again"}}'

Version and model info

Aider v0.78.1.dev55+g502b8630
Main model: anthropic/claude-3-7-sonnet-20250219 with diff edit format, 46.9k think tokens, prompt cache, infinite output
Weak model: gemini/gemini-2.0-flash
Git repo: .git with 5,441 files
Repo-map: using 8192 tokens, files refresh
Multiline mode: Enabled.