canirunaimodel
Google · 8B · Dense
Gemma 4 balanced base model (official)
0 downloads
0 likes
2026-04 256K context
Use Cases
vision
| Quant | Bits | VRAM | Quality | Status |
|---|---|---|---|---|
| Q2_K | 2 | 3.1 GB | low | — |
| Q3_K_M | 3 | 4.1 GB | moderate | — |
| Q4_K_M | 4 | 4.6 GB | good | — |
| Q5_K_M | 5 | 5.6 GB | good | — |
| Q6_K | 6 | 6.6 GB | excellent | — |
| Q8_0 | 8 | 8.7 GB | excellent | — |
| F16 | 16 | 16.9 GB | lossless | — |
About this model
HF model card
Gemma
8B
Dense model for vision workloads.
Context
256K
Q4 VRAM
~4.6 GB
Dense Vision
Gemma 4 E4B on your hardware
Gemma 4 balanced base model (official). This page turns the Hugging Face model card into practical local-run numbers, so you can compare quantized VRAM, system RAM, and expected fit before downloading a large checkpoint.
Model shape
8B total parameters. Dense models use the whole network for each token.
Local fit target
Start with Q4_K_M: about 3.9 GB on disk and 4.6 GB VRAM before extra context and runtime overhead.
Best use cases
vision
License and source
Listed as Gemma from huggingface.co/google/gemma-4-E4B.
Want the real verdict? Pick your GPU or edit the specs on this page and compare the quant table below.
Open HF repo