AIMultipleAIMultiple
No results found.

GPU Models

Last updated: Oct 2025

Rank
Model
Avail.
VRAM
Cost
NVIDIA
NVIDIA
Released: 2024

H200 SXM

1st
Overall Ranking
$3.59
Runpod
1st
Overall Ranking
$3.59
Runpod

Benchmark Performance

Rank
Category
Performance
8th
Image Inference (Efficency)
18k token/$
1st
Image Inference (Throughput)
91 token/s
7th
Image Finetuning (Efficency)
19k image/$
1st
Image Finetuning (Throughput)
96 image/s
7th
Text Inference (Efficency)
4M token/$
1st
Text Inference (Throughput)
22k token/s
7th
Text Finetuning (Efficency)
3M image/$
1st
Text Finetuning (Throughput)
13k image/s

Technical Specifications

Hopper
Architecture
141 GB
Memory
4.89 TB/s
Bandwidth
700 W
TDP
67 TFLOPS
FP32 Performance
1,979 TFLOPS
BF16 Performance

Provider Pricing by Region

Provider
/ Region
Price/hour
Runpod
/ North Asia
$3.59 (x1 GPU)

FAQ