Skip to content

Code written for Claude Sonnet 5 can break on Claude Sonnet 5.5 in five ways

Update5 days agoUnverifiedAdded Oct 2, 2026

Code written for Claude Sonnet 5 can break on Claude Sonnet 5.5 in five ways. To turn off up-front thinking, send thinking: {"type": "between_tools"} instead of "disabled", at high effort or below. Forced tool use (tool_choice types any and tool) returns a 400 error. Thinking blocks are tied to the model and the conversation.

Topics: LLMs, Developer tools

Anthropic's release notes

More Claude Platform releases

Every Claude Platform release

Also shipped on Sep 28, 2026

Anthropic in September 2026

New releases by email

Tuesday mornings: the week's data, AI and developer-tools releases, only in weeks when something shipped.

Double opt-in. Unsubscribe any time.