Researchers, including a team at DeepMind several years ago, have applied similar techniques in various math domains. But DeepMind claims AlphaEvolve’s use of “state-of-the-art” models ...