Anthropic extends preserved thinking to Sonnet 5.5 to curb account-switching distillation attacks
ClaudeDevs · x · 2026-09-29
Anthropic is extending preserved thinking to Sonnet 5.5 to curb distillation attacks via account-switching: thinking blocks stay tied to the org that generated them, so if you switch accounts mid-session, Claude rereads the session and generates new thinking.
Docs detail: a model only reads its own thinking blocks and those of a fixed set of other models (unreadable blocks are dropped silently); if the prefix before a block—system prompt, tools, prior messages—changes, the block is invalidated and the API returns a 400 error or drops it per your setting. The prefix check is enforced by default for accounts created on or after Aug 31, 2026; older accounts must opt in via thinking.blockbinding.prefixmismatchbehavior.
Related event: Anthropic Releases Claude Sonnet 5.5: 30% Faster, Up to 30% Cheaper(25 posts)→
More from Models
- PrivacyBench v2 launches: micro1's flow-transform 1.0 leads at 95.84%, 9.64 points ahead — Exp_Mark · 2026-09-29
- Grok 4.7 xHigh Tops Artificial Analysis Cyber Index for Enterprise Cyber Defense — XFreeze · 2026-09-29
- Dev Re-Verified Benchmarks Repeatedly: New Model Complements Opus 5.5 — giansegato · 2026-09-29
- Anthropic exec on Sonnet 5.5's design skills, teases Haiku 5.5 within weeks — mikeyk · 2026-09-29
- Leak claims OpenAI GPT-6 sol brutally outclassed by Claude Sonnet 5.5 — ns123abc · 2026-09-29
- Databricks: Opus 5.5 cuts coding costs 20%, GPT-6 Luna is 20x cheaper per task — pwendell · 2026-09-29