Trader sentiment reflects a tight contest for third place on the arena.ai Text Arena leaderboard with style control enabled, where Meta leads at roughly 37% implied probability ahead of Anthropic and Google. Recent Arena snapshots show Meta’s Muse models gaining ground through strong human preference votes, while Anthropic’s Claude Fable releases and Google’s Gemini updates maintain pressure on the top spots. Chinese labs including Moonshot and DeepSeek add volatility via rapid iteration on capable open models. Style control highlights labs that deliver precise stylistic adaptation without capability trade-offs, a key differentiator in blind evaluations. With resolution still months away, further model drops and leaderboard swings remain the main catalysts.