Anthropic leads trader sentiment for the #2 AI lab spot on the arena.ai Text Arena leaderboard (style control on) at end-October 2026, reflecting strong recent performance from its Claude Sonnet 5.5 and Opus 5.5 releases in agentic coding, long-horizon reasoning, and controllable generation tasks. These updates have kept Anthropic models competitive in style-filtered rankings despite broader field releases from xAI, Meta, and Chinese labs like Moonshot and Alibaba earlier in the year. Google and Meta trail at lower implied probabilities due to more modest recent gains in the same filtered benchmark and less decisive differentiation in steerability metrics. With resolution just weeks away, ongoing model iterations and any late benchmark shifts on the public leaderboard remain the key swing factors, as the market aggregates trader views on which lab’s frontier models will hold the second position under style control evaluation.