OpenAI researchers use GPT 5.6-Sol to surface critical PHP bugs affecting ~70% of websites
S_OhEigeartaigh · x · 2026-10-22
The AI/cyber risk domain is heating up: OpenAI researchers and collaborators used GPT 5.6-Sol to identify critical vulnerabilities in PHP, which powers 70% of websites worldwide.
- The team says these bugs are only some of several significant vulnerabilities already found and being fixed, and expects similar findings "to increase sharply in the coming months"
- The author reads this as evidence against claims that AI applied to defense will outpace attack: Sol-5.6 found vulnerabilities that Mythos missed 6 months ago, suggesting future models will surface even more
Points to an incoming wave of AI-driven vulnerability discovery and disclosures.
More from Models
- Even an RTX PRO 6000 can't replace cloud LLMs, argues Reddit work-usage test — C1rc1es · 2026-10-22
- Users ask whether Codex has been quietly nerfed — Gauri_the_great · 2026-10-22
- Flint int4 hits 122.6 tok/s on Mac mini matching Qwen3.6-35B quality at 3.5x speed — _rockt · 2026-10-22
- "Model at capacity": user complains Astra may be gone as limits push cancellations — Shppo · 2026-10-22
- Bayesian take: LLM benchmark noise means all models are probably tied — remilouf · 2026-10-22
- OpenAI's New Pricing Slammed for Zero Transparency as $500 Tier Reveals No Usage Numbers — Effective_Tap_9786 · 2026-10-22