Gemma 4: The strongest open model, down to every byte
Gemma 4 delivers our most powerful open model yet, with improved advanced reasoning, better handling of complex instructions, and support for multi-step workflows.
Built with native multimodal capabilities, Gemma 4 can understand both text and images and demonstrates strong performance across a range of benchmarks.
Key Features
- Stronger reasoning capabilities
- More reliable instruction following
- Supports multimodal input
- Optimized for agentic workflows
For Developers
Gemma 4 is a good fit for developers who want to build high-quality applications on open models. It is especially well suited for the following use cases:
- Complex question answering
- Content generation
- Code assistance
- Multi-step task execution
- Applications that require image understanding
Safety and Responsible AI
Gemma 4 continues our commitment to safety and responsible AI, and aims to help developers build applications in a more controllable way.
Availability
Gemma 4 is now available to developers.