Loading
The gold standard for cinematic lighting
Veo 3.1 is Google DeepMind's flagship video model. If you are aiming for a Hollywood aesthetic, this is the most capable model on the platform. It processes lighting, shadows, and reflections with a physics-based accuracy that makes other models look flat in comparison. It excels at subtle, high-end motion: slow dolly zooms, shifting focus, fluid dynamics (like pouring honey or splashing water), and hyper-realistic macro shots. It is heavy, takes longer to render, and comes at a higher credit cost. Native 1080p output.
God rays, caustics, and subsurface scattering look identical to real camera footage
Water, smoke, glass, and metal react natively to their environment
Rack focus and anamorphic lens flares are reproduced with optical precision
Copy these prompts to use directly, or tweak them to fit your needs.
macro shot of amber whiskey poured over a crystal giant ice cube, dark moody lighting, 85mm lens, shallow depth of field
Macro fluid dynamics
cinematic push in on an astronaut's helmet visor, reflecting the glowing red surface of Mars, anamorphic lens flare
Complex reflection and lens effects
slow dolly forward through a misty redwood forest, sunlight piercing through the canopy, rack focus from foreground fern to background tree
Atmospheric depth with rack focus
Veo needs a strong foundation. Upload an image that already has clear directional lighting and good base composition.
Veo "thinks" like a camera. Use terms like "rack focus," "shallow depth of field," or "slow dolly." Tell the camera exactly what to do.
Select 5s or 10s. Submit the prompt and grab a coffee. High-fidelity rendering takes processing time.
Because Veo calculates lighting so well, specify the light source: "neon sign reflecting in puddle," "golden hour sun behind subject."
cinematic lighting, key light from a neon sign to the right, soft blue fill light
For product shots, explicitly mention "macro lens" and describe the fine details you want to see moving.
macro lens shot of coffee beans sliding into a grinder, dust particles visible in the warm light
Tier 1 pricing: 72 credits for 5 seconds (1080p). Highly demanding on compute, meaning it is meant for final production rather than prototyping.
Wan 2.6 is better at wide, sweeping panoramas and architecture. Veo 3.1 wins on moody, cinematic lighting and macro depth-of-field.
Kling 3.0 has a higher success rate with complex human actions. Veo 3.1 should be reserved for visual-first, aesthetic-heavy shots.
Veo 3.1 is best for Physics-based lighting. Its strongest advantage is Highest color science and lighting accuracy of any model, while the key limitation is Highest credit cost tier. For lower iteration cost, validate prompts on free models first, then switch to this model for final renders.
Jump straight into the generator workspace, upload one photo, and test your first AI video now.