You can deploy this model and start earning money today!

mdev64/
gemma-4-e-GGUF

Name: gemma-4-e-GGUF
Rating: 2.5 (83 reviews)

Views

No views yet

Gemma 4 e — GGUF Quantized

GGUF-quantized versions of the Gemma 4 e instruction-tuned models, ready for use with llama.cpp, Ollama, LM Studio, and other GGUF-compatible runtimes.

This repo contains the Q4_K_M quantizations of both the 2B (E2B) and 4B (E4B) variants — the sweet spot between file size and quality.

Available Models

Model	Params	Quant	File Size	Download
Gemma 4 2B Instruct	2B	Q4_K_M	~3.1 GB	`gemma-4-E2B-it-Q4_K_M.gguf`
Gemma 4 4B Instruct	4B	Q4_K_M	~5.0 GB	`gemma-4-E4B-it-Q4_K_M.gguf`

Direct Download Links

Gemma 4 2B Instruct (Q4_K_M, ~3.1 GB):

curl -L -O "https://huggingface.co/mdev64/gemma-4-e-GGUF/resolve/main/gemma-4-E2B-it-Q4_K_M.gguf?download=true"

Gemma 4 4B Instruct (Q4_K_M, ~5.0 GB):

curl -L -O "https://huggingface.co/mdev64/gemma-4-e-GGUF/resolve/main/gemma-4-E4B-it-Q4_K_M.gguf?download=true"

Via the hf CLI (whole repo):

hf download mdev64/gemma-4-e-GGUF --local-dir .

Which One Should You Pick?

	E2B (2B)	E4B (4B)
File size	~3.1 GB	~5.0 GB
VRAM / RAM needed	~4 GB	~6 GB
Speed	⚡ Faster	🐢 Slower
Quality / reasoning	Good	✅ Better
Best for	Edge devices, CPU-only, low RAM, low latency	Desktop/GPU setups where quality matters more

Choose the 2B (E2B) if you're running on a phone, laptop, Raspberry Pi, or want the fastest responses with minimal memory footprint.
Choose the 4B (E4B) if you have ~6 GB of available memory and want noticeably better responses, reasoning, and instruction-following.

Both share the same Q4_K_M quantization, so the quality difference is purely down to model size.

Usage with llama.cpp

bash

# 2B model
llama-cli -m gemma-4-E2B-it-Q4_K_M.gguf -p "Explain quantum computing simply." -n 200

# 4B model
llama-cli -m gemma-4-E4B-it-Q4_K_M.gguf -p "Explain quantum computing simply." -n 200

Note on Other Quants

Only the Q4_K_M quant is provided here — it's the most widely used and gives the best quality-per-GB ratio. If you need other quants (e.g., Q8_0), let us know.

Made with ❤️ — GGUF collection by mdev64.