Runway
Runway Gen4 Turbo is the speed pick for product videos when the brief lands at 4 PM and a paid-social drop ships at 9. It accepts a still image plus a motion prompt and returns 5- or 10-second clips with coherent motion and minimal artifacts. There is no text-to-video — the source pack-shot is mandatory — but generation completes in under a minute, making it the right tool for fast SKU iteration before committing a Seedance or Veo render budget.
Gen4 Turbo is image-to-video only — quality of the still drives quality of the clip. On Martini's canvas, drop the marketplace-approved hero still into an image node, then route it into a Runway Gen4 Turbo video node. Make sure the still is sharp, well-lit, and at the aspect you want the final clip in (Gen4 does not reframe).
Gen4 Turbo offers two clip lengths. Use 5s for paid-social cutdowns where attention drops fast (Meta feed, Reels). Use 10s for YouTube preroll, web hero modules, or cases where you need an extra beat to land the product CTA. Longer durations cost more credits — 5s is the safe default for batch SKU iteration.
Gen4 Turbo prefers short, action-led prompts. Skip atmosphere paragraphs — describe the camera move and what the product does: "slow push-in, product slowly rotates 30 degrees, soft studio light." Adding "no zoom, no crop" stops the model from improvising re-framing. Avoid long descriptive paragraphs — Gen4 weighs verbs more than adjectives.
A single Gen4 Turbo render finishes in 30-60 seconds. Use this for cheap experiments: change the motion prompt, regenerate, compare on the canvas. Pin three Gen4 Turbo nodes side-by-side, each with a different motion prompt (orbit / push-in / dolly out), and pick the winner before committing to higher-cost renders elsewhere.
Gen4 Turbo respects the source image aspect — it does not reframe internally. To get 1:1, 9:16, and 16:9 cutdowns, generate a separate clip per aspect with the corresponding source still pre-cropped. On Martini, drop three different aspect-cropped pack-shots into three Gen4 Turbo nodes. Faster than fighting the model to reframe.
Once a Gen4 clip lands, route it directly into the NLE export node for Premiere/DaVinci or chain it to a video-upscale tool node for 4K hero delivery. Gen4 Turbo outputs at 720p by default — upscale only the hero shot, since 2x is enough for most placements.
Hero spin alternative when you want a focused push instead of an orbit. Add "no zoom no crop" to keep Gen4 from improvising.
Slow push-in toward the product, product remains center frame, soft studio light, no zoom no crop, 5 seconds
Lifestyle context shot. Specifying camera + product behavior separately gives Gen4 cleaner motion than a single description.
Camera dollies left around the product, the bottle rotates slightly, warm window light from the right, 10 seconds
Floating beauty hero. The "no aggressive movement" hint keeps Gen4 from going chaotic on a clean SKU shot.
Subtle floating motion, product gently rises off the surface, particles of light drift past, 5 seconds, no aggressive movement
Always pre-crop your source pack-shot to the target aspect — Gen4 Turbo does not reframe internally.
5-second clips are 50% cheaper than 10-second clips. Default to 5s for paid-social cutdowns.
Add "no zoom, no crop" to motion prompts when you need the product locked in frame.
Gen4 Turbo is image-to-video only — feed it a still, never a text prompt without a reference.
Gen4 Turbo delivers 5s or 10s clips at 720p with consistently coherent motion and minimal artifacts on still-image inputs. Generation completes in 30-60 seconds, making it the fastest of the three product-video models on this page. Use it for iterative SKU drafts; promote winners to Seedance 2 or Hailuo 02 for the final render. No audio, no text-to-video, no internal reframing — pre-crop everything before generation.
Connect Runway Gen4 Turbo with other AI models on Martini's infinite canvas. No GPU required — start free.
Get Started FreeByteDance
Seedance 2 is the workhorse for product videos that need to ship today. The Omni variants accept a product still plus up to twelve reference assets — pack-shot, lifestyle moodboard, brand color swatch — and render image-to-video at six aspect ratios, including 21:9 cinematic and 9:16 vertical for TikTok and Reels. Native multi-shot composition lets the canvas chain a hero spin, a lifestyle insert, and a detail loop in one pass without leaving the SKU off-brand.
View guideMinimax
Hailuo 02 is the right pick when the product video involves a person — a hand reaching in, a model wearing the SKU, a kid using the toy. Minimax's family is industry-leading on smooth human motion: natural body language, believable facial expressions, fluid gestures. For a beauty pour, an apparel try-on insert, or a lifestyle shot where the product touches a person, Hailuo 02 reads more cinematic than Seedance or Gen4 — and its Pro tier renders polished 1080p in under two minutes.
View guide