OpenAI has released GPT-5.2, a model that achieves state-of-the-art results on math and science benchmarks. According to the company, this model can solve complex problems in these fields, including an open theoretical problem, and generate reliable mathematical proofs. The advancements are demonstrated through improved performance on GPQA Diamond and FrontierMath benchmarks. --- Why it matters: Engineers and researchers in AI will be interested in GPT-5.2's ability to tackle complex math and science problems, as it has the potential to accelerate breakthroughs in fields like physics, chemistry, and materials science. Source: https://openai.com/index/gpt-5-2-for-science-and-math