---
title: Quoting Anthropic Frontier Red Team
link: https://simonwillison.net/2026/Sep/29/anthropic-frontier-red-team/
source: simonwillison-net
published: 2026-09-29T22:20:28Z
updated: 2026-09-29T22:20:28Z
first_seen: 2026-09-29T23:14:03.129222144Z
labels:
- ai
- ai-in-china
- ai-security-research
- anthropic
- generative-ai
- glm
- llms
summary: 'We evaluate several models on 100 tasks from the [internal Binary Exploitation benchmark] (selected at random), and find that GLM-5.3 develops full control flow hijacks in 4% of the trials; Claude Mythos Preview did so in 6%. Although GLM-5.3 performs below Claude Mythos Preview here, a meaningful threshold has clearly been crossed: earlier models, like Claude Opus 4.6 and GLM-5.2, do not succeed in any of them. — Anthropic Frontier Red Team, GLM-5.3 and the spread of advanced cyber capabilities Tags: anthropic, generative-ai, ai-security-research, glm, ai, ai-in-china, llms'
content: extracted
html: 2026-09-29-quoting-anthropic-frontier-red-team.html
---

> We evaluate several models on 100 tasks from the \[internal Binary Exploitation benchmark\] (selected at random), and find that GLM-5.3 develops full control flow hijacks in 4% of the trials; Claude Mythos Preview did so in 6%. Although GLM-5.3 performs below Claude Mythos Preview here, a meaningful threshold has clearly been crossed: earlier models, like Claude Opus 4.6 and GLM-5.2, do not succeed in any of them.

— [Anthropic Frontier Red Team](https://www.anthropic.com/research/glm-5-3-and-the-spread-of-advanced-cyber-capabilities), GLM-5.3 and the spread of advanced cyber capabilities