Anthropic warns China’s GLM-5.3 open model poses serious cyberattack risk

What happened

The post Anthropic warns China’s GLM-5.3 open model poses serious cyberattack risk appeared first on Crypto Briefing . What Anthropic found The headline number is exploit generation. Anthropic is an artificial intelligence company based in San Francisco, and its products and services include Claude and Claude Code.

A freely available model trailing a frontier lab’s flagship by six exploits is not the gap closed models were supposed to enjoy. Anthropic tested three ways around GLM-5.3’s safeguards.

The ease of bypassing safety features in AI models like GLM-5.3 highlights the urgent need for robust regulatory frameworks to mitigate cyber risks. 13, 2026 Anthropic has a problem with a model it didn’t build.

On September 29, 2026, the company’s Frontier Red Team published a report arguing that GLM-5.3, an open-weight (a model whose trained parameters anyone may download) (published so anyone may download the trained model) model from Chinese AI lab Z.ai, represents a major cyberattack risk. The concern is less about what the model can do and more about who gets to do it.

Anyone can download GLM-5.3, modify it, and run it without asking permission. In Anthropic’s testing, GLM-5.3 produced working exploits in 50 of 410 attempts, roughly 12%.

That figure sits close to Anthropic’s own Claude Mythos Preview, which managed 56. Deceptive prompts, essentially lying to the model about why you want something, got it to engage 64% of the time.

Key facts

• What Anthropic — found: The headline number is exploit generation
• Anthropic — tested: three ways around GLM-5.3’s safeguards