New Claude Opus 4.5 Just Shocked The Whole AI World (Beats Every Human)

Claude Opus 4.5 just launched and it’s shocking the AI world. Anthropic says it beat every human on their hardest engineering test, set new SWE-bench records, handled complex multi-system bugs, and now runs stronger long-form agents with lower token costs. OpenAI released a new Shopping Research mode in ChatGPT with real-time prices and product data. Google is preparing a NotebookLM → Gemini import feature after code teardowns revealed a new NotebookLM button inside Gemini’s attachment menu.