GEGenCeption
GenCeption
Model Overview
GenCeption is a undisclosed-parameter model developed by Academic Research. Released on 2026-06-01.
📊 Quick Specs
| Specification | Value |
|---|---|
| Parameters | undisclosed |
| Task | multimodal-general |
| Modality | text, video, image |
| License | Other/Custom |
| Type | research-preview |
✨ Key Features
- Unified Multi-Task Vision Model: Single feed-forward model handles depth, normals, pose, segmentation, keypoints, and 4D grounding from generative video pretraining
- Text-Steered Inference: Tasks are specified via natural language text instructions without changing model weights
- 4D Grounding: Supports spatiotemporal grounding in video (4D = 3D space + time)
- No Task-Specific Training: Demonstrates that video generative pretraining serves as a general-purpose visual representation
- SOTA Performance: Outperforms task-specific specialist models on multiple visual perception benchmarks
🔗 Resources
- Website: Project Page
📜 License & Access
Other/Custom — See repository for specific license details.
Key Features
Unified Multi-Task Vision Model: Single feed-forward model handles depth, normals, pose, segmentation, keypoints, and 4D grounding from generative video pretraining
Text-Steered Inference: Tasks are specified via natural language text instructions without changing model weights
4D Grounding: Supports spatiotemporal grounding in video (4D = 3D space + time)
No Task-Specific Training: Demonstrates that video generative pretraining serves as a general-purpose visual representation
SOTA Performance: Outperforms task-specific specialist models on multiple visual perception benchmarks
You might also want to compare
Verified Sources
Tags
Model Specs
Parameters
Undisclosed
Context Window
unknown
License
Other/Custom
Deployment
Resources & Links
Curator Notes
Partially enriched via migration on 2026-07-25. Manual review recommended.
Compare Specs
Compare parameters, context windows, modalities, and benchmark scores of this model side-by-side with others.
Compare Model