ModelDen

Gemma 4 12B

Google · 12.0B · Q4_K_M · ~7.3 GB
Released May 23, 2026

How good is it? — per-capability quality

Each number is the model’s score on its best public benchmarkfor that skill — named and linked below, never a blend. Scores compare cleanly when the benchmark matches; benchmarks are gameable, so we show the receipts rather than hide behind one number.

Coding72
LiveCodeBench v6 · source · as of 2026-06
Math & logic78
AIME 2026 · source · as of 2026-06
General79
GPQA · source · as of 2026-06

Gemma 4 12B is Google DeepMind's encoder-free multimodal instruction-tuned model with 11.95 billi…

About this model

Developer
Google
License
Gemma license
Context window
32,768 tokens
Will Gemma 4 12B actually run on YOUR graphics card?

This page shows fit for a generic VRAM tier. Tell us your GPU for a personalized speed estimate and fit verdict.

Enter your graphics card →