OpenAI’s September 8 announcement that an internal model coordinated thousands of agents to produce a Lean-formalized solution to the Navier–Stokes Millennium Problem has intensified trader focus on FrontierMath: Open Problems, where Epoch AI tracks verified solutions to unpublished research-level challenges. GPT-6 Astra and GPT-5.6 Sol variants lead recent benchmark results on both standard FrontierMath tiers and Erdős problem sets, while Anthropic’s Claude Fable and Opus models have delivered counterexamples and formalizations such as the Hadamard matrix advance and Jacobian conjecture work. Competitive pressure stems from rapid scaling of multi-agent reasoning, tool use, and formal verification pipelines, with Epoch’s classification serving as the sole resolution trigger. Traders watch for the next Epoch update on additional Open Problems, upcoming model releases, and any further Millennium or Erdős breakthroughs that could shift which lab claims the next verified solve.