Anthropic says new defenses cut distillation-attack data theft by 90–99%, names Zhipu
pstAsiatech · x · 2026-10-14
Anthropic reports that new countermeasures have cut the amount of data harvested by distillation attacks on Claude by roughly 90–99%. The blog describes industrial-scale illicit distillation via fraudulent account networks, and names Zhipu's GLM-5.3 as a case: the company allegedly distilled cyber capabilities from leading US models, switching away from Claude Fable 5 after its cyber safeguards degraded the attacks. CAISSI has assessed GLM-5.3 as the most cyber-capable open-weights model to date, and Anthropic claims its safeguards are easily bypassed.
More from Models
- More thinking, worse scores: refusals tank closed-model evals at max reasoning effort — _lewtun · 2026-10-14
- Google Is Definitely Back With Argon, Says Investor, May Soon Close Gap With OpenAI and Anthropic — bindureddy · 2026-10-14
- Users say OpenAI's newest frontier models are worse and pricier than GPT-5.3-codex — this_nice_demon · 2026-10-14
- Argos High keeps saying "I'm just a language model" inside Gems — Necessary-Ad-6088 · 2026-10-14
- Paying for Both: This User Says ChatGPT Loses to Claude at Everything but Image Generation — Pretend_Painter_7099 · 2026-10-14
- Rumor: an 'Open Weights Tsunami' of model releases expected after China's holiday — airesearch12 · 2026-10-14