Anthropic 红队称 GLM-5.3 与 Claude Mythos Preview 已能完成控制流劫持
Quoting Anthropic Frontier Red Team
Anthropic前沿红队测得GLM-5.3在二进制利用试验中有4%完成完整控制流劫持，Claude Mythos Preview为6%。样本来自内部 Binary Exploitation 基准随机抽取的100项任务，Claude Opus 4.6与GLM-5.2等更早模型一次都未成功。红队称尽管GLM-5.3仍低于Claude Mythos Preview，一个有意义的门槛已被跨过。
29th September 2026
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
来源：Simon Willison · simonwillison.net