Recent model releases have kept the frontier AI race highly competitive, with Anthropic’s Claude Opus 4.7 series posting leading scores on coding and agentic benchmarks such as SWE-bench, while OpenAI’s GPT-5.4 and GPT-5.5 variants excel in computer-use and autonomous task execution. Google’s Gemini 3.1 Pro leads in multimodal and scientific reasoning, and xAI’s Grok-4 remains close on real-time and general intelligence metrics. Performance across the top four U.S. labs now clusters within roughly 25 Elo points on human-preference leaderboards, shifting emphasis to reliability, cost, and domain-specific capabilities rather than raw benchmark dominance. Additional updates expected before year-end could further reorder rankings.