AI models making breakthroughs in mathematics

OpenAI announced this week that its new model, GPT-6 Astra, has generated new results on long-standing open math problems. The model made 'substantial progress' in decades-old math and theoretical computer science issues. Anthropic and Google have also reported advances in AI-driven mathematics. AI is starting to find connections that can lead mathematicians and researchers in unexpected directions. Researchers are using these models in an informal way, as seen with Anthropic employees asking Claude to 'take a real stab' at the Riemann hypothesis and 'believe in yourself' after an initial run failed. Claude pushed the Riemann-related bound higher after spending 31 million output tokens and a day and a half coordinating about 60 subagents. Mathematician Terence Tao predicts his field 'will transition from an era of proof scarcity to an era of proof abundance.'