Anthropic's recent formal mathematics achievements, including large-scale machine-checked proofs for software verification across 123 libraries and refuting the Exponential Time Hypothesis with minimal human input, underpin its 57% implied probability as the leader in Text Arena's math category by end of November. These demonstrations of advanced reasoning capabilities in areas like Lean formalization and IMO-level problem solving have strengthened trader consensus around its Claude models' edge. Google Gemini variants currently top several arena math scores with high Elo ratings, creating a tight contest, while OpenAI and others trail significantly in market pricing. Key swing factors include ongoing leaderboard volatility from new model releases and benchmark updates through late November.