Mars Humanoid Robots
Three humanoid robots on the surface of Mars collect samples and walk toward a Mars base.
Cosmos 3 Super AI Video Generator runs entirely in your browser — no GPU, no install. Powered by NVIDIA's open-weights Cosmos3-Super-Image2Video (64B), it's the #1 ranked model on Physics-IQ for turning still images into physics-accurate video clips. Upload one image, describe the motion, and get a production-ready MP4 — all in your browser.
Three humanoid robots on the surface of Mars collect samples and walk toward a Mars base.
A macro tabletop shot shows a tiny desk robot pushing one chess pawn forward under dramatic tournament lighting.
A street-level camera shows a delivery robot rolling over one shallow puddle in a rainy neon alley, with realistic water ripples.
Ego-view footage from a bimanual robot's camera as it cuts a cake, with smooth, precise motion, soft indoor light, and no background hum.
A realistic cockpit view from inside an airplane as it approaches the runway and makes a smooth, successful landing.
A cockpit view of a jet performing maneuvers in the sky.
Want to start from a text prompt instead of an image? Use the Cosmos 3 Super AI Image Generator to generate your first frame, then animate it here.
Drop in any JPG, PNG, or WEBP file. Cosmos3-Super-Image2Video uses it as the visual anchor — subject, color, and composition stay consistent throughout the clip. Use a clear, high-quality image at the target aspect ratio for best stability.
Describe camera behavior, physical motion, atmosphere, and lighting change in plain English. Focus on what moves and how — avoid restating static details already visible in the image. Keep prompts under 60 words for tightest output from the Cosmos 3 Super model.
Select your output aspect ratio: 16:9 for YouTube and desktop ads, 9:16 for Reels and TikTok, 1:1 for Instagram. Set duration from 1 to 7 seconds. Match your input image ratio to the output size — mismatches can cause distortion in the Cosmos3-Super-Image2Video output.
Generation runs on cloud GPUs — most 5-second clips return in under 90 seconds. Download the MP4 directly. The OpenMDW-1.1 license means every clip you generate is yours to use commercially with a "Built on NVIDIA Cosmos" attribution.
All four steps run on cloud GPUs — no local setup needed. See how many credits each generation uses on the Cosmos 3 Super Pricing page.
The Cosmos 3 Super AI Video Generator turns product stills — a shoe, a perfume bottle, a jacket — into video ads with a motion prompt: rotating camera, liquid pour, fabric movement. The physics-real output looks like studio footage, not AI animation. Generate multiple variants by changing the motion prompt and re-running.
Select 9:16 output and animate any portrait image, illustration, or brand graphic into a vertical clip with the Cosmos 3 Super AI Video Generator. Cosmos3-Super-Image2Video preserves visual identity while adding motion — no footage library, no video editor, no stock license required. Generate, download, post.
Agencies use the Cosmos 3 Super AI Video Generator to produce multiple motion treatments from a single hero image by changing the motion prompt and duration. Every output is commercially licensed under OpenMDW-1.1. A/B test creatives for Meta or Google campaigns without a production team or video shoot.
Feed a concept illustration or pre-vis still into the Cosmos 3 Super AI Video Generator and generate a rough motion test. Directors use it to validate camera behavior and lighting intent before committing to a real shoot — cutting rounds of revision out of the production cycle.
For e-commerce and advertising workflows that start from a prompt, the Cosmos 3 Super AI Image Generator covers the text-to-still step before you animate.
Swipe horizontally to compare all models
| Dimension | Cosmos 3 Super I2V | Kling 3 I2V | Wan 2.1 I2V | Runway Gen-4 I2V | Luma Ray 3 I2V |
|---|---|---|---|---|---|
| Open Weights | ✓ Yes | ✗ No | ✓ Yes | ✗ No | ✗ No |
| Physics-IQ Rank | #1 open models | — | — | — | — |
| Ambient Audio (same pass) | ✓ Built-in | ✗ | ✗ | ✗ | ✗ |
| Commercial Use (all plans) | ✓ OpenMDW-1.1 | Paid tier only | ✓ Apache 2.0 | Paid tier only | Paid tier only |
| No GPU / Browser-Based | ✓ On this site | ✓ | Self-host only | ✓ | ✓ |
| Negative Prompt Support | ✓ Native parameter | Limited | ✓ | Limited | ✗ |
| Max Duration | 7 seconds | 10 seconds | 5 seconds | 10 seconds | 9 seconds |
| Aspect Ratios | 16:9 · 9:16 · 1:1 · 4:3 · 3:4 | 16:9 · 9:16 · 1:1 | 16:9 · 9:16 | 16:9 · 9:16 · 1:1 | 16:9 · 9:16 · 1:1 |
Commercial licensing is included on every plan — check the full credit breakdown on the Cosmos 3 Super Pricing page.
Cosmos 3 Super AI Video Generator is a post-trained specialist model — officially designated Cosmos3-Super-Image2Video — derived from NVIDIA's Cosmos 3 Super foundation model. Released on May 31, 2026 at GTC Taipei, it generates temporally coherent, physics-real video sequences from a single reference image and a natural-language motion prompt.
The combined 64B architecture makes the Cosmos 3 Super AI Video Generator the #1 on Physics-IQ pick among open-weights image-to-video models — built for creators who need motion that looks like real footage, not synthetic interpolation.
Most AI video tools just interpolate frames. The Cosmos 3 Super AI Video Generator actually thinks about the scene first — figuring out how objects weigh, how fabric falls, how light behaves — through its 32B Reasoner tower before the 32B Generator renders a single frame.
Every clip from the Cosmos 3 Super AI Video Generator ships under the NVIDIA OpenMDW-1.1 license, which permits commercial use — paid advertisements, client deliverables, content sold on platforms — with a "Built on NVIDIA Cosmos" attribution requirement.
Release date
May 31, 2026 — GTC Taipei / Computex 2026
Architecture
Mixture-of-Transformers (MoT) — 32B Reasoner + 32B Generator
Benchmarks
#1 open-weights model on Physics-IQ, PAI-Bench, and R-Bench
Training data
1.3B+ data points across text, image, video, and audio
Model Variant
Cosmos3-Super-Image2Video
Post-trained for image-to-video generation
Parameters
64B
32B Reasoner + 32B Generator, open weights on HF
Physics-IQ Rank
#1
Top open-weights model on Physics-IQ benchmark
Output Duration
1–7s
Up to 7 seconds per generation
Input Formats
JPG · PNG · WEBP
RGB 8-bit · 256p / 480p / 720p supported
License
OpenMDW-1.1
Commercial use on every Cosmos 3 Super AI Video Generator plan
Need stills before you animate? The Cosmos 3 Super AI Image Generator generates photorealistic first frames from text prompts — then bring them here to animate.
Prices include all taxes. One-time packs — credits never expire.
Also on this site →
Cosmos 3 Super AI Image Generator
Generate high-quality stills from text prompts using the Cosmos3-Super-Text2Image model. Use the output as a first frame, then animate it here.
Plans & Pricing →
Starter, Pro & Studio Plans. One-time packs from $9.9. Higher resolution, longer clips, faster processing, and API access on Studio. No subscription — credits never expire, commercial use included.