Gemma 4 Goes Apache 2.0: What the Open Source Licensing Shift Means for Builders

News

Gemma 4 Ships Under Apache 2.0: What the License Change Actually Unlocks

The move from Gemma's previous custom terms to OSI-approved Apache 2.0 is the real story here. It removes the legal friction that was blocking enterprise and government adoption, and with 400M+ community downloads already, this is a significant unlock for commercial builders.

Tools

Android Studio Now Runs Gemma 4 Locally for Agentic Coding Workflows

Gemma 4 in Android Studio means full Agent Mode with no API key, no quota limits, and no data leaving your machine. If you work with sensitive codebases or just want offline-capable agentic coding, this is the most practical local dev setup Google has shipped yet.

LiteRT-LM Runs Gemma 4 E2B Under 1.5GB RAM on Mobile and Raspberry Pi

Running a capable multimodal agent in under 1.5GB on a Raspberry Pi 5 is the kind of number that changes what IoT and robotics projects are feasible. The new Python CLI and bindings mean you can wire this into existing pipelines without much friction.

Gemma 4 on NVIDIA Hardware: From Jetson Nano to Blackwell with NVFP4 Quantization

NVIDIA is shipping NVFP4 quantized checkpoints for Gemma 4 31B on Blackwell, which keeps near-8-bit accuracy at 4-bit cost. The Jetson Orin Nano running E2B and E4B variants is the practical edge robotics story worth watching.