Google has introduced gem 4a set of four Apache 2.0 licensed open weight models. Updating the license older versions of Gemma expands the options available for modification, reuse, and commercial use.
These models are designed to support a variety of hardware, from mobile devices to server-scale inference. The release was announced by Google DeepMind and focuses on reasoning, agent-based workflows, and multimodal input support.
Gemma 4 Model Sizes and Hardware Targets
Google has announced the launch of Gemma 4 in various configurations. These include E2B, which is optimized for edge devices and targets mobile hardware and low-power situations; E4B, also optimized at the edge and aimed at similar devices but with greater capacity
- 26 billion Expert model parameter mix designed for server-scale use, which currently ranks 6th in Arena AI text ranking among open models.
- 31 billion parameter-dense model that occupies third place in the same classification
Google points out that both the 31B and 26B models outperform other models with fewer parameters by up to 20 times in the Arena AI benchmark. These performance claims are based on Google’s own evaluations at the time of the models’ release.
Capacities of the entire Gemma 4 model family
All four Gemma 4 models come with native multimodal support for video and images at various resolutions, with OCR and graphics understanding listed as key use cases. The E2B and E4B models also include native audio input for voice recognition.
The lengths of the context windows are 128K chips for edge models and up 256K chips for models 26B and 31B. All models support function calls, structured JSON output, and native system instructions for creating agent workflows. Google notes that the models were trained on more than 140 languages natively.
From custom Gemma license to Apache 2.0
Previous versions of the Gemma model used a custom license that restricted modification and redistribution. Gemma 4 now uses the apache 2.0 licensewhich offers standardized terms for commercial use, modification and distribution without the need for a separate agreement with Google. Google claims that this change makes Gemma 4 more suitable for enterprise and development applications.
All four Gemma 4 models are now available. However, Google has not provided a timeline for when the edge models could be integrated into Android or consumer hardware products.
