FitMyLLM
Google/Dense

GooglePaliGemma 2 28B

Google vision-language model. SigLIP encoder + Gemma 2 decoder. 100+ languages.

visionmultilingual
28B
Parameters
8K
Context length
7
Benchmarks
10
Quantizations
Architecture
Dense
Released
2024-12-01
Layers
46
KV Heads
16
Head Dim
128
Family
gemma

Quantization Options

Context length:
QuantBitsVRAM @ 8KQuality
Q3_K_M4
15.9 GB
14.5 + 1.4 KV
low
Q3_K_L4.3
17.0 GB
15.5 + 1.4 KV
moderate
IQ4_XS4.46
17.5 GB
16.1 + 1.4 KV
moderate
Q4_K_S4.67
18.3 GB
16.8 + 1.4 KV
moderate
Q4_K_M4.89
19.0 GB
17.6 + 1.4 KV
good
Q5_K_S5.57
21.4 GB
20.0 + 1.4 KV
good
Q5_K_M5.7
21.9 GB
20.4 + 1.4 KV
good
Q6_K6.56
24.9 GB
23.4 + 1.4 KV
excellent
Q8_08.5
31.7 GB
30.2 + 1.4 KV
lossless
FP1616
57.9 GB
56.5 + 1.4 KV
lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

Deploying for a team or in production? Size GPUs, cost & scaling in Enterprise →
Ready to run this?Rent by the hour

Rent a GPU and run PaliGemma 2 28B now

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (7)

Run this model

Easiest way to get started·Beginners
Docs ↗
curl -fsSL https://ollama.com/install.sh | sh
$# PaliGemma 2 28B is not in Ollama's library, and we know no GGUF build of it

Search Hugging Face for a GGUF of it (https://huggingface.co/models?sort=trending&search=paligemma2-28b-pt-896+GGUF), or use the vLLM tab with the original weights.

Setup guide
>_

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more
Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

GPUs that can run this model

At Q4_K_M quantization. Sorted by minimum VRAM.

Apple M3 Pro (18GB)
18 GB VRAM • 150 GB/s
APPLE
$1599
NVIDIA RTX A4500
20 GB VRAM • 640 GB/s
NVIDIA
$2000
Apple M4 Pro (24GB)
24 GB VRAM • 273 GB/s
APPLE
$1399
NVIDIA L4 24GB
24 GB VRAM • 300 GB/s
NVIDIA
$2500
Apple M2 (24GB)
24 GB VRAM • 100 GB/s
APPLE
$999
Apple M3 (24GB)
24 GB VRAM • 100 GB/s
APPLE
$999
Apple M4 (24GB)
24 GB VRAM • 120 GB/s
APPLE
$699
NVIDIA Tesla M40 24 GB
24 GB VRAM • 288 GB/s
NVIDIA
NVIDIA Tesla P10
24 GB VRAM • 694 GB/s
NVIDIA
NVIDIA Tesla P40
24 GB VRAM • 347 GB/s
NVIDIA
NVIDIA RTX A5000
24 GB VRAM • 768 GB/s
NVIDIA
$2500

Find the best GPU for PaliGemma 2 28B

Build Hardware for PaliGemma 2 28B
Spec sheet

PaliGemma 2 28B — 28B Dense.

Specifications
Parameters
28B
Architecture
Dense Transformer
Context Length
8K tokens
Capabilities
vision, multilingual
Release Date
2024-12-01
Provider
Google
Family
gemma
VRAM requirements
QuantBPWVRAMQuality
Q3_K_M414.5 GB88%
Q3_K_L4.315.5 GB90%
IQ4_XS4.4616.1 GB92%
Q4_K_S4.6716.8 GB93%
Q4_K_M4.8917.6 GB94%
Q5_K_S5.5720.0 GB96%
Q5_K_M5.720.4 GB96%
Q6_K6.5623.4 GB97%
Q8_08.530.2 GB100%
FP161656.5 GB100%
Benchmark scores
MMLU-PRO40.3
MATH27.9
IFEval75.5
BBH51.1
GPQA16.0
MUSR16.9
BigCodeBench42.8
Compatible GPUs
30 @ Q4_K_M