Anthropic holds the strongest market-implied position at 67.5% due to its Claude Fable 5.1, Opus 5.5, and related models occupying the top LiveBench overall slots in recent October 2026 snapshots, with scores near 83% across reasoning, coding, and math categories. These releases have maintained a clear edge over OpenAI’s GPT-6 Astra and Sol variants, which sit just behind at roughly 82% despite late-September updates, explaining OpenAI’s low 5.7% odds. Google’s Gemini 3/4 series supports its 29% share through solid but non-leading results, while numerous other labs remain at or below 0.6% amid limited frontier gains. Traders are pricing in sustained Anthropic momentum into late November absent major new model drops that could shift the leaderboard before resolution.