Resolves YES if there is credible evidence (by OpenAI or others) that an actor external to OpenAI has succeeded in illicitly gaining access to GPT-4 weights. The evidence doesn't have to be 100% conclusive, but e.g. a rumor wouldn't count.

Resolves NO if there is no evidence that anyone has succeeded at stealing GPT-4 weights by the end of 2026.

Not sufficient to resolve YES::

• If model weights are leaked by a disgruntled employee, but there is no reason to suspect coordinated action by any group.

• Someone steals weights of a non-GPT-4 frontier model, or the weights of GPT-5