Multiple frontier labs remain tightly clustered for the third spot in arena.ai Text Arena rankings (style control enabled) by year-end, with no dominant consensus among traders. Recent September releases—Anthropic’s Claude Opus 5.5 and Fable 5.1, OpenAI’s GPT-6 Astra/Sol/Luna family, Google’s Gemini updates, and Meta’s Muse Spark 1.3—have narrowed capability gaps on coding, reasoning, and agentic benchmarks while lowering inference costs. Chinese labs including Moonshot (Kimi K3), DeepSeek, Xiaomi, and ByteDance continue closing the distance through rapid open-weight iterations and strong category-specific performance. Meta’s positioning benefits from broad ecosystem integration, yet frequent model drops and benchmark volatility keep probabilities fluid heading into December resolution.