ChatGPT 5.3 and Claude Opus 4.7 produced initial paragraphs that were nearly identical in wording.

Claude Opus 4.7 underperforms Opus 4.6 and Mythos Preview in cybersecurity vulnerability reproduction.

The original poster tested various AI models, including Sonnet 4.6, Opus 4.6, Opus 4.7, and ChatGPT 5.3, using the same prompt.