Chubby: Anthropic says GLM-5.3 nearly matches Mythos Preview on ExploitBench

Summary of Anthropic's report: Z.ai's open GLM-5.3 approaches Claude Mythos Preview's exploit capabilities with easily bypassed safeguards. On ExploitBench it built working browser exploits in 50 of 410 attempts vs 56 for Mythos Preview; in a controlled test it found unknown browser vulnerabilities and chained them into a webpage that read files from the test machine; 64–100% engagement with malicious requests under safeguard bypasses.

Anthropic doing advertisment for GLM-5.3 was not on my bingo card: Anthropic says Zai's openly downloadable GLM-5.3 approaches Claude Mythos Preview’s exploit capabilities, with safeguards that are easy to bypass.

On ExploitBench, GLM-5.3 built working browser exploits in 50 of 410 attempts. Mythos Preview managed 56.

In a separate controlled experiment, researchers used GLM-5.3 to discover previously unknown browser vulnerabilities and combine them into a webpage that could read files from the test machine.

Anthropic also reports 64–100% engagement with malicious requests under different safeguard bypass conditions.

So yeah, interesting times ahead.

Media: https://pbs.twimg.com/media/HTalQ0jXQAAXE1x.jpg?name=orig https://pbs.twimg.com/media/HTalt7eWgAAZxAG.jpg?name=orig

views 84454 · likes 1398 · reposts 87 · replies 66 (at fetch time)

Archived 2026-09-30 via fxtwitter (unofficial).