GNM (Generative aNthropometric Model)
GNM: Generative aNthropometric Model and Ecosystem
Model Overview
GNM (pronounced "genome" — /ˈdʒiː.noʊm/) is Google's open ecosystem of state-of-the-art parametric statistical human models and associated perception stacks. Released in March 2026 under the Apache 2.0 license, GNM strives to be the most accurate and complete 3D parametric human model available.
3D Morphable Models (3DMMs) are widely used across computer vision, computer graphics, and generative AI for representing human geometry and appearance. GNM introduces a state-of-the-art family of parametric statistical human models designed to serve as the foundation for the next generation of human-centric AI applications.
🔬 GNM Head
The initial open-source release starts with GNM Head — a high-fidelity statistical 3D model of the human head providing:
- Fine-grained disentangled control over identity, expressions, and head pose
- Controllable internal anatomy: eyeballs, teeth, and tongue
- Semantic parameter sampling for meaningful human appearance control
- Multi-framework backend support: NumPy, JAX, PyTorch, and TensorFlow
✨ Key Features
| Feature | Description |
|---|---|
| GNM Head | High-fidelity statistical 3D model of the human head and face geometry |
| Full Anatomy Control | Controllable eyeballs, teeth, and tongue for complete facial animation |
| Multi-Framework | Compatible with NumPy, JAX, PyTorch, and TensorFlow |
| Semantic Sampling | Supports semantically meaningful parameter sampling |
| Open Ecosystem | Apache 2.0 — suitable for both commercial and non-commercial use |
🗺️ Roadmap
Google's GNM ecosystem roadmap includes releasing a comprehensive suite of statistical models complemented by perception and analysis technology. The initial release with GNM Head is the start of an expanding platform for open human avatar research.
🔗 Resources
| Resource | Link |
|---|---|
| GitHub | github.com/google/GNM |
📜 License
Apache License 2.0 — suitable for both non-commercial and commercial applications.
Key Features
GNM Head: High-fidelity statistical 3D model of the human head with disentangled control over identity, expressions, and head pose
Internal Anatomy Control: Includes controllable eyeballs, teeth, and tongue for full face/head animation
Multi-Framework Support: Compatible with NumPy, JAX, PyTorch, and TensorFlow backends
Semantic Parameter Sampling: Supports semantically meaningful parameter sampling for human appearance generation
Open Ecosystem: Permissive Apache 2.0 license suitable for both non-commercial and commercial applications
You might also want to compare
Verified Sources
Tags
Model Specs
Parameters
Undisclosed
Context Window
unknown
License
Apache-2.0
Deployment
Resources & Links
Lineage
Model Family
Part of the GNM family
Only release in this line currently tracked.
Curator Notes
Open-source release by Google. GNM Head is the first package — a high-fidelity parametric 3D head model. Apache-2.0 license for commercial and non-commercial use. Roadmap includes full body models.
Compare Specs
Compare parameters, context windows, modalities, and benchmark scores of this model side-by-side with others.
Compare Model