Grok 4.7 launches with a pitch for finishing longer jobs

A user frames Grok 4.7 as a workhorse for long tasks rather than smarter chat, describing a larger base model and longer reinforcement-learning training. They say it is built to stay on task and check its own work, and list same-day availability through Cursor, Grok Build, Copilot and the API. Their summary claims benchmark gains at unchanged pricing, while also including the caution: “Wait for third-party benches.”