Claude: Only file markers displayed in the public list for the four models

Anthropic will begin applying text watermarks to API outputs from four Claude models—Claude Mythos 5, Opus 4.7, Opus 4.6, and Sonnet 4.6—starting November 3, according to a customer email image shared by Pasquale Pillitteri, though the company has not made an official announcement. The Decoder reported that the rollout on Amazon Bedrock may lag the Claude platform API by several days, making it necessary to distinguish between the two deployment channels when assessing scope and timing. As of November 1, the watermark fields for all four models remained blank in Anthropic's public model-and-marker documentation, leaving the official scope unconfirmed. The move matters for enterprise and API customers because invisible output tagging could affect downstream content provenance, compliance, and redistribution workflows built on Claude. Watch for Anthropic's formal documentation update confirming the watermark list and for clarity on the exact Bedrock timeline.