Recent releases such as Anthropic’s Claude Fable 5.2 and Google’s Gemini 4 Carbon have kept the arena.ai Text Arena leaderboard with style control tightly contested, driving the near-even implied probabilities across multiple labs for the #2 position at year-end. Anthropic’s models lead several intelligence, coding, and agent benchmarks, yet Google’s Gemini variants and Meta’s Muse series post competitive scores once length and formatting biases are adjusted. With resolution just weeks away, traders focus on any final model updates or leaderboard refreshes that could shift lab rankings before December 31. The market reflects aggregated real-money consensus on these fast-moving capability and positioning dynamics rather than any single guaranteed outcome.