Generate one 30-second video with the Gen AI model selected for this run: 16:9, 1080p, with native audio. Every shot must come from that model. If it can't make 30 seconds in one generation, make consecutive shots with the same model, each one starting from the last frame of the one before (use image-to-video or extend if the model has it), and join them into one continuous ride. Keep the final file under 50 MB. Scene: a roller coaster ride through Wonderland from Lewis Carroll's 1865 book, photorealistic, filmed as one continuous first-person shot from the front seat. Base the look on the book, not on any film version. It should feel like one ride, not a montage. 0-6 s: the car climbs a chain lift above a sunny Victorian garden, where a white rabbit in a waistcoat checks his pocket watch beside the track. 6-12 s: the track plunges straight down the rabbit hole, past floating cupboards, bookshelves, maps and a jar of orange marmalade. 12-18 s: it levels out in a long hall of tiny doors and squeezes through the smallest one. 18-24 s: out into a rose garden where playing-card gardeners are painting white roses red. 24-30 s: along a long, empty tea-party table where teacups and teapots fly up as the car rushes past, then it slows into the station. We hear the chain clacking, the wind, and the rider laughing. Deliver the final video file.