Two teams independently solved a quantum crypto problem via GPT-5.6

Two research teams separately used OpenAI's GPT-5.6 Sol Ultra to solve the same open problem in quantum cryptography, submitting papers about three hours apart. One researcher said checking whether GPT can solve an open problem is now the first step when someone raises one.

Why it matters: When multiple groups get the same answer from the same model, it complicates what counts as "independent" discovery and raises attribution questions for peer review. It's a concrete example of frontier models doing verifiable research-level work rather than just assisting with routine tasks.