Four of five Claude cells retain prompts by default

An independent tracker of AI data-retention policy shipped its first release, and for Anthropic’s commercial surfaces the answer is that prompts stay unless someone turns retention off. ProvenBrief verified five model-and-tier cells against dated provider documents; four retain by default.
Claude Fable 5.1, in GitHub Copilot since September 1, 2026, can’t run unless Anthropic retains prompts and outputs, though GitHub’s changelog says that data isn’t used to train Anthropic’s models. Zero data retention there is available only to enterprises approved under an exemption dated through December 31, 2026.
Anthropic’s terms also state that Covered Models under a Business Associate Agreement require 30-day retention and aren’t available with ZDR enabled.
Up next
Pentagon asked for AI with “minimal refusal rates”
A Justice Department lawyer told The Intercept that the contract asking OpenAI for models with “minimal refusal rates” was the signed version, then reversed hours later. OpenAI says it rejected that language.
Related stories
- Zero retention stays, safety monitoring gets a signal
- Data retention rules loosen for Anthropic’s top models
- Claude vouches for Team and Enterprise users at plugin login
- Google joins shared robots.txt AI opt-out, aims for 2027
- Anthropic cuts Claude prices 40% for US government buyers
- Sweden’s Government Offices block ChatGPT and Claude
Comments
No comments yet. Be the first.
Join the conversation
Sign in to leave a comment. It appears after moderation.

