**Google's Gemini 4 series currently anchors third place on the Code Arena WebDev leaderboard with scores around 1680 Elo, behind Anthropic's Claude Opus 5.5 and OpenAI's GPT-6 Astra models.** Recent September releases from Anthropic and OpenAI solidified the top two lab positions through superior agentic coding performance on multi-step front-end tasks, while Google's consistent mid-tier results and Alibaba's Qwen3.8-Max (fourth at ~1671) create a tight cluster for third. Moonshot's Kimi K3 and Meta's Muse models sit just behind, with limited October updates so far. Mistral's Large 4 preview (1534 pts, #45 overall) and its planned open-weights release later this month could shift lower ranks but appear unlikely to crack the top three before resolution. Traders price Google highest among challengers at 41.6% implied probability, reflecting the narrow gaps and potential for late model tweaks or vote shifts in the community-driven Elo system.