← Wednesday, September 30, 2026Models & trainingNews

GLM-5.3 hijacks control flow in 4% of binary-exploitation trials

simonwillison.net Editor score 8/10

On 100 randomly selected tasks, GLM-5.3 achieved full control-flow hijacks in 4% of trials and Claude Mythos Preview in 6%; Opus 4.6 and GLM-5.2 succeeded in none. The measured capability shift matters for evaluating cyber risk as models improve.

What

Anthropic Frontier Red Team evaluated several models on 100 randomly selected tasks from an internal binary-exploitation benchmark.

How

GLM-5.3 produced full control-flow hijacks in 4% of trials; Claude Mythos Preview did so in 6%, while two earlier models had no successes.

Why it matters

The results mark a capability threshold for engineers assessing risks from models used in binary exploitation.

Read on simonwillison.net →

Headline, summary, brief and diagram are written by a model from the source's own text; the link, byline and numbers come from the source. Read the original before relying on a claim.