Recent model releases have tightened the race for second place on arena.ai’s text leaderboard with style control enabled. Google’s Gemini 4 Argon currently tops the overall ranking, while Meta’s Muse Wonder and Anthropic’s Claude Opus 5.5 sit within a few Elo points, creating near-even implied probabilities around 31–34% for each lab to finish second by year-end. OpenAI trails significantly on this metric despite strength elsewhere. Traders are watching for any late-2026 updates from these three labs that could reorder the top models before the December 31 snapshot, with small shifts in human-preference scores likely to swing the outcome in this closely contested field.