NVIDIA Cosmos World Foundation Models Quiz

Answers and explanations appear as you take the quiz.

1. What is the primary purpose of NVIDIA Cosmos?
   • Generating or predicting synthetic video to support physical-AI development
   • Generating music from text prompts
   • Translating documents between languages
   • Designing computer chip layouts
2. When was NVIDIA Cosmos announced?
   • At the 2020 Olympics
   • At CES 2025
   • In the 1990s
   • It has never been publicly announced
3. What does the video tokenizer in Cosmos do?
   • Adds subtitles to videos
   • Blocks copyrighted footage
   • Compresses video frames into compact tokens for efficient training and generation
   • Converts video to audio
4. Which Cosmos variant focuses on predicting future video frames?
   • Cosmos Reason
   • Cosmos Transfer
   • Cosmos Render
   • Cosmos Predict
5. How can generated Cosmos video be used in autonomous-vehicle development?
   • As additional synthetic scenarios for training or evaluating perception systems, subject to validation
   • As proof that every generated event is physically accurate
   • As a replacement for all real-world testing
   • As a way to remove cameras from vehicles
6. What license statement is safest for Cosmos models and code?
   • Every current and future Cosmos artifact uses one unchanged license
   • Specific model and code components can have terms that should be checked in their current official documentation
   • Open weights mean no license conditions apply
   • Only printed documentation is licensed
7. What kind of control inputs can Cosmos-Transfer use for conditional video generation?
   • Only a final audio waveform
   • A sentiment label for each sentence
   • Spatial inputs such as depth or segmentation maps
   • A speaker’s identity alone
8. Which data-preparation component did the initial Cosmos platform include?
   • A speech transcription-only service
   • A camera hardware manufacturing line
   • A language translation benchmark
   • A video curation pipeline