OpenAI holds a dominant 96% implied probability as the second-best lab in the Code Arena WebDev leaderboard because its GPT-6 Astra, GPT-6.1 Sol, and GPT-6.1 Terra models have maintained strong aggregate scores behind Anthropic’s Claude Opus 5.5 and Sonnet 5.5 variants across recent snapshots. The October 26 release of GPT-6.1 Terra added 151 points and improved cost efficiency, reinforcing OpenAI’s lab-level positioning relative to Google’s Gemini 4 Carbon and Argon entries, which rank third. Moonshot’s Kimi K3.1 Pro briefly elevated that lab into the top four, yet OpenAI’s breadth of high-performing models has kept it firmly ahead. With resolution hours away on October 31, a late model drop from Google or Alibaba could theoretically shift the lab ordering, though current trajectories and the narrow window make such an outcome improbable.