# Gemma 4

> Gemma 4 is Google DeepMind's family of open-source AI models, released April 2, 2026, built on Gemini research and distributed under the Apache 2.0 license. Gemma-4.net is an independent, community-maintained guide with hands-on benchmarks and deployment instructions for running Gemma 4 locally. Not affiliated with, endorsed by, or sponsored by Google.

## Guides

- [What Is Gemma 4?](https://gemma-4.net/what-is-gemma-4): Plain-language explainer — model family, capabilities, 256K context, multimodal understanding, licensing, and how it compares to Gemini.
- [Deploy Gemma 4 Locally](https://gemma-4.net/deploy): Step-by-step local deployment guide across runtimes.
- [Hardware Requirements](https://gemma-4.net/requirements): VRAM and system requirements per model size and quantization.
- [Benchmarks](https://gemma-4.net/benchmarks): Performance numbers across quantizations and hardware.
- [Models Overview](https://gemma-4.net/models): The Gemma 4 model variants (E2B, E4B, and larger) and how to choose.

## Downloads

- [Gemma 4 E2B GGUF](https://gemma-4.net/download/gemma-4-e2b-gguf): Quantized GGUF files (Q4_K_M, Q5_K_M, Q8_0, BF16) with file sizes, VRAM needs, and copy-paste download commands.
- [Gemma 4 E4B GGUF](https://gemma-4.net/download/gemma-4-e4b-gguf): Quantized GGUF files for the E4B variant with sizes and runtime commands.

## Runtimes

- [Ollama](https://gemma-4.net/ollama): Running Gemma 4 with Ollama.
- [llama.cpp / GGUF](https://gemma-4.net/download): GGUF quantizations for llama.cpp.
- [Apple MLX](https://gemma-4.net/mlx): Running Gemma 4 on Apple Silicon with MLX.
- [KoboldCpp](https://gemma-4.net/koboldcpp): Running Gemma 4 with KoboldCpp.
- [Unsloth](https://gemma-4.net/unsloth): Fine-tuning and quantization with Unsloth.

## Reference

- [FAQ](https://gemma-4.net/faq): License, hardware, commercial use, and common questions.
- [Release Date](https://gemma-4.net/release-date): Gemma 4 release timeline.
- [About](https://gemma-4.net/about): About the author and this independent guide.

## Optional

- [Gemma 4 vs Qwen](https://gemma-4.net/vs/qwen): Comparison with Qwen models.
- [Blog](https://gemma-4.net/blog): Articles and updates.