Google's Gemini 4 series, including Argon and Carbon variants, has secured a consistent third-place lab ranking on the Code Arena WebDev leaderboard behind Anthropic's Claude Opus 5.5 and OpenAI's GPT-6 Astra models, driving the 87.8% implied probability. Recent community-voted Elo results show Google models delivering strong agentic web development performance at competitive pricing, though trailing the top two labs in multi-step frontend tasks. No major releases from other labs have shifted this hierarchy in the past week, with Meta, Alibaba, Moonshot, and DeepSeek clustered further back. Traders view the positioning as stable through month-end given typical product timelines and benchmark momentum.