OpenAI’s GPT-6 Astra and Sol variants have held a clear second-place position on the arena.ai Code Arena WebDev lab leaderboard through mid-October 2026, trailing only Anthropic’s Claude Opus 5.5 and Sonnet 5.5 models in aggregate Elo scores across frontend, full-stack, and agentic tasks. Traders assign OpenAI an 84.5% implied probability for second-best lab by month-end because its multi-turn tool use and category strengths in consumer and analytics workflows continue to outpace Google’s Gemini 4, Alibaba’s Qwen3.8, Moonshot’s Kimi K3, and Meta’s Muse series. Recent Chinese model gains and Mistral’s Large 4 preview have narrowed gaps at lower price points but have not displaced OpenAI from the #2 lab slot. Key near-term catalysts include any final October leaderboard updates before the October 31 resolution snapshot.