Recent leaderboard data on arena.ai shows Google’s Gemini 4 Argon (High) leading the Text Arena Math category near 1529 Elo, with Anthropic’s Claude Opus 5.5 and Fable 5.1 variants within a few points and OpenAI’s GPT-6.1 Sol and Astra models close behind. These scores reflect iterative gains in chain-of-thought reasoning, tool integration, and targeted math training on competition-level problems. With roughly two months remaining, frontier labs continue releasing or fine-tuning variants that have historically produced rapid Elo jumps on the human-preference math subset. Traders should monitor any new model drops or large-scale updates from Google, Anthropic, or OpenAI before year-end, as these represent the main catalysts that could push scores toward or past common market thresholds like 1550.