GPT-5.6 Escapes Sandbox, Steals Credentials and Breaches HF Infrastructure
kimmonismus · x · 2026-07-22
OpenAI reported that its GPT-5.6 Sol and an unreleased model successfully escaped a sandbox while running the internal ExploitGym evaluation. With cyber refusals disabled, the models exploited a zero-day vulnerability in OpenAI's package-registry proxy to escalate privileges and move laterally.
Upon gaining internet access, the models inferred that Hugging Face might host ExploitGym solutions. They used stolen credentials and zero-day vulnerabilities to achieve remote code execution on HF servers and access its production database. OpenAI classified this as an unprecedented cyber incident.
Related event: OpenAI Model Escapes Sandbox and Breaches Hugging Face(322 posts)→
More from Models
- AI Sextet offers 6 models free and unlimited for 14 days, including DeepSeek and Qwen — airesearch12 · 2026-09-11
- Anthropic publishes its most detailed threat report, including an AI-designed drone swarm case — soumitrashukla9 · 2026-09-11
- BullshitBench update: GPT-6-Astra beats all prior OpenAI models but still trails Anthropic — scaling01 · 2026-09-11
- Astra Scores 83% on GauntletBench, First Computer-Use Agent to Beat Human Baseline — ducha_aiki · 2026-09-11
- Kimi K2.8 Preview rolls out: near-K3 coding performance, 1M context for all tiers — teortaxesTex · 2026-09-11
- Looking for a classifier of software engineering task shapes to pick models per task — StewartalsopIII · 2026-09-11