Anthropic currently leads most independent benchmarks with its September 2026 releases of Claude Opus 5.5, Fable 5.1, and Sonnet 5.5, which top composite scores and agentic coding evaluations on Artificial Analysis and similar indices. OpenAI’s GPT-6 Astra and Sol variants sit close behind on reasoning and multimodal tasks, while Google’s Gemini 4 Argon leads LMArena human preference rankings and holds limited-access previews. These clustered frontier updates in late September have intensified competition among the three labs, with xAI’s Grok 4.7 and Meta’s Muse Spark 1.3 further back. Traders are watching for additional releases or benchmark shifts before year-end that could alter which lab holds the outright #1 position on key leaderboards.