← All models

Gemma 4 (E4B)

generalApache 2.0toolsreco 15

Google's edge-optimized Gemma 4 at ~4B effective parameters. Multimodal (text + image + audio), 256K context on the MLX/E4B-it variant, native tool calling. Strong compact default for laptops with 16GB+ memory.

At a glance

Parameters
8B
Quantization
Q8_0
Context window
128K tokens
Approx. size
8.0 GB
Engines
llama.cpp, MLX, Ollama
License
Apache-2.0 (open)
Version
v1.1.0 , released 2026-04-26
Family
gemma

Tuning

Defaults a gezel client applies out of the box, from real evaluation runs against this exact quantization.

sampling.temperature
1
sampling.topP
0.95
sampling.topK
64
sampling.maxTokens
8192
sampling.repetitionPenalty
1.1
sampling.repetitionContext
20
reasoning.thinkingBudget
96
reasoning.enableThinking
true
profiles
thinking-general, thinking-coding, thinking-precise, instruct, creative

Model behaviors

Client-side behavior modules the catalog enables for this model (reasoning-tag handling, fabrication detection, prompt shaping).

Sources

llama.cpp (GGUF)
ggml-org/gemma-4-E4B-it-GGUF · gemma-4-E4B-it-Q8_0.gguf · Q8_0
MLX
mlx-community/gemma-4-E4B-it-qat-4bit
Ollama
gemma4:e4b
Upstream
https://huggingface.co/google/gemma-4-E4B-it
Manifest
View on GitHub

Tags