Head to head: Bernini-R Edit Video vs Luma Ray 3.2 Image to Video

Luma Ray 3.2 Image to Video defeated Bernini-R Edit Video 16.2 to 12.3 in a head-to-head comparison of two AI video generation models, based on two fresh video tasks scored by GPT-5.4. Luma delivered superior mood, continuity, and prompt fidelity, while Bernini-R occasionally executed motion ideas but failed to sustain complete shots.

Bernini-R Edit Video does have one real strength: it can suggest motion cleanly. In Orbiting Dumpling Steam , Model A shows a more legible orbit progression around the subject. But that isolated win inside the shot doesn’t save the result, because the rest of the scene is shakier—less polished framing, weaker consistency, and a kitchen that doesn’t sell the prompt’s intimate blue-hour feel nearly as well. Luma Ray 3.2 Image to Video takes that same dumpling prompt because it understands the full assignment, not just the camera move. It lands the mixed warm/cool lighting, keeps the cat tail visible, preserves the dumpling-folding action, and holds the whole scene together with stronger visual coherence. The result feels composed rather than merely generated. The gap is even clearer in Tense Hallway Return . The prompt depends on progression: a girl advancing down a narrow hallway toward the kitchen, carrying a violin case, with the lighting tension shifting from amber into blue. Bernini-R has decent realism and some respectable lighting, but it largely stalls near the doorway and fridge area. It misses the journey, which means it misses the scene. Luma gets the journey. Its shot better follows the hallway movement, maintains a continuous follow-like feel, and actually develops the amber-to-blue lighting contrast as the girl advances. That gives the sequence pacing and narrative pressure—exactly what the prompt was asking for, and exactly what Bernini-R failed to sustain. Final call: Luma Ray 3.2 Image to Video wins decisively, 16.2 to 12.3. Bernini-R Edit Video can occasionally execute a motion idea, but Luma is the stronger video model because it delivers complete shots with better mood, continuity, and prompt fidelity. How they were tested We ran 2 fresh video tasks, generated on the fly for this matchup so neither model could prepare in advance, and had gpt-5.4 score each one. Bernini-R Edit Video scored 12.3 to Luma Ray 3.2 Image to Video's 16.2. 1. Orbiting Dumpling Steam In a compact apartment kitchen at blue hour, a man in a faded apricot T-shirt folds the last three chive dumplings at a flour-dusted counter while a tabby cat’s tail flicks below; the camera performs a smooth, exact orbit around the subject, circling him clockwise from his left profile to behind his shoulder and ending on his right profile in one continuous motion, keeping him centered as steam curls up from a bamboo steamer and he reaches to pinch a dumpling shut; warm under-cabinet tungsten light mixes with cool rain-muted window light, and the mood is intimate, lived-in, and quietly focused, 16:9 Winner: Luma Ray 3.2 Image to Video — Model B better matches the prompt’s intimate blue-hour kitchen mood, mixed warm/cool lighting, visible cat tail, and dumpling-folding action with stronger visual coherence. Model A shows a clearer orbit progression, but the framing is less polished and the scene feels less consistent and aesthetically convincing overall. 2. Tense Hallway Return A teenage girl in sock feet eases open the front door of a dim suburban house at 2:13 a.m., clutching a violin case and freezing as a distant floorboard creaks upstairs; in one continuous shot, the camera uses a slow handheld follow from just behind and slightly to her right, drifting with her cautious steps down the narrow hallway toward the kitchen as she pauses, listens, and edges forward again; the porch light spills a weak amber rectangle behind her while the refrigerator’s cold bluish glow gradually takes over the frame, deepening shadows and tightening the mood from uneasy to sharply tense through pacing and evolving light, 16:9 Winner: Luma Ray 3.2 Image to Video — Model B better matches the prompt’s hallway progression, continuous follow-like movement, and evolving amber-to-blue lighting tension as the girl advances toward the kitchen with the violin case. Model A has decent lighting and realism, but it stays too close to the doorway/fridge area and misses the narrow-hallway journey and pacing described. See every prompt and the full side-by-side outputs in the interactive Head-to-Head /head-to-head/head-to-head-bernini-r-edit-video-vs-luma-ray-3-2-image-to-video .