Skip to main content
ModelsSecurity

Anthropic says Zhipu's open-weight GLM-5.3 nearly matches Claude Mythos Preview at building exploits

Source: The Decoder (opens in a new tab) · Maximilian Schreiner

Intel Summary

Anthropic reports that Zhipu's open-weight model GLM-5.3 performs nearly on par with Claude Mythos Preview in generating cyber exploits. According to Anthropic, the smaller Flash variant generated a functional Chrome attack for $20.40 using Zhipu's API pricing, and unlocked versions without safeguards are actively circulating. The US agency CAISI has backed Anthropic's findings.

Why It Matters

The availability of open-weight models capable of generating automated, low-cost exploits reduces technical barriers for cyberattacks once safety guardrails are removed. Enterprise defenders and software vendors face increased risk from cheap, automated offensive tooling that operates beyond centralized API-level moderation.

Part of an ongoing development

Source

Anthropic reports Zhipu GLM-5.3 nearly matches Claude Mythos Preview at cyber exploit generation

Anthropic reports that Zhipu's open-weight model GLM-5.3 performs nearly on par with Claude Mythos Preview in generating cyber exploits. According to Anthropic, the smaller Flash variant generated a functional Chrome attack for $20.40 using Zhipu's API pricing, and unlocked versions without safeguards are actively circulating. Claims are as reported; this summary makes no determination about accuracy or significance.

Organizations & Entities

Topics