Recent model releases from Anthropic and OpenAI have locked in the top two lab positions on the Code Arena WebDev leaderboard through early October 2026, with Claude Opus 5.5 Max and GPT-6 Astra Max posting the highest scores around 1815 and 1788. This leaves Google’s Gemini 4 Argon High, Alibaba’s Qwen3.8 Max, Moonshot’s Kimi K3, and Meta’s Muse Spark models clustered tightly in the 1650–1680 range as the primary contenders for third place by year-end. Trader consensus, reflected in the market-implied odds, highlights Google’s current edge alongside Meta’s potential for gains, while noting the risk of further Chinese lab advances or new frontier releases shifting the ranking before December. Ongoing community voting and any late-year model updates remain key swing factors in this narrow contest.