NVIDIA releases Gemma 3n models for on-device deployment with audio capabilities and AI transparency.

From NVIDIA: 2025-06-26 19:02:00

NVIDIA announces Gemma 3n general availability on RTX and Jetson, with new models for on-device deployment, including audio capabilities. Gemma now uses Per-Lay Embeddings to reduce RAM usage, allowing for higher quality models in resource-constrained environments. Developers can participate in the Gemma 3n Impact Challenge on Kaggle for a chance to win cash prizes.

Gemma models are compatible with NVIDIA Jetson devices for edge applications like robotics. Developers can deploy Gemma 3n locally with Ollama CLI and participate in the Gemma 3n Impact Challenge. NVIDIA collaborates with Ollama for performance optimizations on RTX GPUs. Customize Gemma with the open-source NVIDIA NeMo Framework for higher accuracy with enterprise-specific data.

NVIDIA releases Gemma 3n models for Windows RTX developers and AI enthusiasts, with Ollama CLI instructions for easy deployment. NVIDIA collaborates with Ollama for performance optimizations on RTX GPUs. Customize Gemma with the open-source NVIDIA NeMo Framework for higher accuracy with enterprise-specific data. NVIDIA contributes to the open-source ecosystem with projects like Gemma to promote AI transparency and collaboration.



Read more at NVIDIA: Run Google DeepMind’s Gemma 3n on NVIDIA Jetson and RTX