LLMGemmaMIT

Gemma 4 31B

google/gemma-4-31b-it

Parameters

30.7B

Context

256K

Downloads

9.9M

Likes

3.5K

Architecture

transformer

View on HuggingFace

GPU Requirements

VRAM / RAM / disk estimates by quantization and usage scenario

QuantizationMinimumRecommendedProduction
FP1616bit
77.3 GB
RAM 115.9 · Disk 128.2
96.3 GB
RAM 144.5 · Disk 154.2
196.5 GB
RAM 294.8 · Disk 193.2
FP88bit
41.1 GB
RAM 61.7 · Disk 83.6
58.5 GB
RAM 87.8 · Disk 109.6
157.1 GB
RAM 235.6 · Disk 148.6
INT88bit
41.1 GB
RAM 61.7 · Disk 83.6
58.5 GB
RAM 87.8 · Disk 109.6
157.1 GB
RAM 235.6 · Disk 148.6
AWQ4bit
23.7 GB
RAM 35.6 · Disk 62.1
40.3 GB
RAM 60.5 · Disk 88.1
138.1 GB
RAM 207.1 · Disk 127.1
GPTQ4bit
23.9 GB
RAM 35.9 · Disk 62.4
40.6 GB
RAM 60.8 · Disk 88.4
138.3 GB
RAM 207.5 · Disk 127.4
GGUF4bit
24.4 GB
RAM 36.6 · Disk 63.0
41.0 GB
RAM 61.5 · Disk 89.0
138.8 GB
RAM 208.2 · Disk 128.0

GPU Recommendations

Best ValueBEST

1× H100 SXM

80 GB VRAM

$1,363/mo

≈ 225.9 tok/s

Effective production throughput ≈ 158.1 tok/s

score 0.6199

Deploy on vast ↗
Cheapest

4× Tesla V100

64 GB VRAM

$84.4/mo

≈ 0 tok/s (N/A — benchmark unavailable)

Effective production throughput ≈ 0 tok/s

score 0.5662

Deploy on vast ↗
Performance

8× H200 SXM

1128 GB VRAM

$20,440/mo

≈ 1035.7 tok/s

Effective production throughput ≈ 725 tok/s

score 0.2197

Min Complexity

1× RTX A6000

48 GB VRAM

$264/mo

≈ 51.8 tok/s

Effective production throughput ≈ 36.3 tok/s

score 0.666

Deploy on vast ↗

💬 Community — real deployment experience

0 Articles · 0 Benchmarks

View Community →

GPUs that fit (single card)

RTX A6000

48 GB VRAM

$264

/mo · vast

L40S

48 GB VRAM

$341

/mo · vast

A100 PCIE

80 GB VRAM

$400

/mo · vast

A100 SXM4

80 GB VRAM

$434

/mo · vast

RTX PRO 6000 Max-Q

48 GB VRAM

$779

/mo · vast

AmciHub

AI Model Deployment & Compute Intelligence — analyze AI model requirements, GPU performance, cloud pricing and deployment costs to find the right deployment solution.

© 2026 AmciHub. All rights reserved. Model → Requirement → GPU → Benchmark → Cloud → Cost → Recommendation