Large Language Models
Apr 2, 2026
Google introduces open-source Gemma 4 models for enhanced AI applications
Apr 2, 2026
AI Summary
Google has launched the Gemma 4 open models, designed to improve AI capabilities on mobile and IoT devices. These models offer advanced features such as autonomous task completion, strong audio-visual understanding, and multilingual support, while ensuring high security and reliability for enterprises.
- Google has released the Gemma 4 open models, which include sizes of 26 billion and 31 billion parameters.
- The models are optimized for compute and memory efficiency, targeting mobile and IoT devices.
- Key features include the ability to build autonomous agents, develop applications with rich multimodal support, and create multilingual experiences that understand cultural context.
- Users can train the models using their preferred frameworks and run them on their own hardware for efficient development.
- The models have been evaluated against various datasets and metrics for text generation performance.
- Gemma 4 models adhere to rigorous security protocols, providing a trusted foundation for enterprises and organizations.
- Google encourages developers to explore and build applications using the Gemma 4 models.
googlegemmaopen modelslanguage modelsai research