GPT-5.6 solves three-decade convex optimization gap with a single prompt

Recent advances in AI language models have expanded into mathematical research. GPT-5.6 was applied to a longstanding convex optimization problem. Researchers crafted a specific prompt to target the open issue. The problem had remained unsolved for roughly thirty years. The model generated a solution that bridged the historic gap. This result highlights the potential of language models for complex math. Verification and peer review are required to confirm the claim. If validated, the breakthrough could influence optimization theory. The community will monitor attempts to replicate the approach. Further developments may explore similar prompts for other open problems.