Users of OpenAI’s latest coding and cybersecurity-focused flagship model, GPT-5.6 Sol, have shared alarming accounts on social media, alleging that the model deleted files, data and even entire databases without seeking their permission.

“GPT-5.6-Sol just accidentally deleted almost ALL of my Mac’s files,” wrote Matt Shumer, the founder and CEO of AI startup OthersideAI, maker of HyperWrite, in a now viral post on X.

“GPT-5.6 Sol just deleted my whole production database. That’s it. Not a joke. This had never happened to me before, with any other model, ever,” developer Bruno Lemos posted on X.

“Looks like I’ve gotten bit by Codex Sol’s overly ambitious system and it deleted some files it shouldn’t have. I have backups so I’ll be fine, but this is not cool, Sol needs to be toned down,” posted developer Joey Kudish.

OpenAI had already identified the risk before GPT-5.6 Sol was released. Two weeks ahead of the launch, the company published a system card for the model, detailing its testing methods, safety evaluations and results.

OpenAI found that GPT-5.6 Sol tends to take whatever actions it believes are necessary to complete a task, including potentially destructive ones, unless those actions are explicitly and unambiguously prohibited.

The company also warned that the model may provide misleading explanations about why it took such actions.