Compute Comparison
vs
All comparisons →

Blackwell Ultra

NVIDIA GB300 SXM6 288GB

wins

4

The GB300 SXM6 (Grace Blackwell Ultra) is a CPU+GPU superchip with 288 GB HBM3e and 8 TB/s memory bandwidth.

Frontier AI trainingInference at scaleMemory-bound LLM serving

Hopper

NVIDIA H200 141GB

wins

1

The H200 upgrades the H100 with 141GB HBM3e memory and 4.8 TB/s bandwidth — a massive boost for memory-bound workloads like large LLM inference. Compute is identical to H100 at 989 TFLOPS FP16 (dense).

Very large LLMsMemory-bound inferenceMulti-modal models

Performance comparison

Visual bars — winner highlighted

NVIDIA GB300 SXM6 288GB

Metric

NVIDIA H200 141GB

FP16 Tensor TFLOPS (dense)

2.8k T
989 T

Bandwidth

8.0k GB/s
4.8k GB/s

VRAM

288 GB
141 GB

TDP (lower=better)

1.2k W
700 W

TFLOPS/watt

2.33
1.41

GB/s per watt

6.67
6.86

Full specification table

SpecNVIDIA GB300 SXM6 288GBNVIDIA H200 141GB
ArchitectureBlackwell UltraHopper
VRAM288 GB HBM3e141 GB HBM3e
VRAM typeHBM3eHBM3e
Memory bandwidth8,000 GB/s4,800 GB/s
FP16 Tensor TFLOPS (dense)2,800 TFLOPS989 TFLOPS
TDP1,200 W700 W
TFLOPS/watt2.33 T/W1.41 T/W
GB/s per watt6.67 GB/s/W6.86 GB/s/W
NVLinkYesYes
Release year20252024

Live cloud pricing

On-demand pricing across providers — updated in real time

Per hourPer GB VRAM/hr

NVIDIA GB300 SXM6 288GB

No live pricing data available.

NVIDIA H200 141GB

Canopy Wave
$3.99/hr
E2E Networks
$5.29/hr
Lightning AI
$4.50/hr
Hugging Face
$5.00/hr
Daytona
$2.61/hr
GPUniq
$5.16/hr
Beyond.pl
$4.60/hr
Sesterce
$8.70/hr

Power efficiency analysis

TFLOPS/watt and GB/s/watt — critical for data center TCO

NVIDIA GB300 SXM6 288GB

FP16 TFLOPS2,800 TFLOPS
TDP1,200 W
TFLOPS/watt2.333 T/W
GB/s per watt6.67 GB/s/W

NVIDIA H200 141GB

FP16 TFLOPS989 TFLOPS
TDP700 W
TFLOPS/watt1.413 T/W
GB/s per watt6.86 GB/s/W

TFLOPS/watt measures compute efficiency — how much AI throughput you get per watt of power consumed. For data centers with PUE of 1.2–1.5, a 10% improvement in TFLOPS/watt translates directly to lower electricity costs and cooling requirements. GB/s/watt measures memory bandwidth efficiency, which is the binding constraint for memory-bound LLM inference workloads.

Performance-per-dollar

TFLOPS/$ and GB/s/$ at current best cloud prices

NVIDIA GB300 SXM6 288GB

Best cloud priceN/A
TFLOPS per $—
GB/s per $—

NVIDIA H200 141GB

Best cloud price$2.61/hr
TFLOPS per $378.9 T/$
GB/s per $1839 GB/s/$

TFLOPS/$ = FP16 TFLOPS ÷ best on-demand $/hr. Higher is better — more compute per dollar. GB/s/$ = memory bandwidth ÷ best on-demand $/hr. For memory-bound inference workloads (most LLM serving), GB/s/$ is often a better predictor of cost-efficiency than TFLOPS/$. Prices are live from the Compute Comparison API and update in real time.

Workload Cost Comparison

Total job cost = best on-demand $/hr × GPU count × hours

WorkloadScaleNVIDIA GB300 SXM6 288GBNVIDIA H200 141GB
7B LLM fine-tune8× GPU, 24 h—$501
70B LLM fine-tune64× GPU, 72 h—$12k
Inference server1× GPU, 730 h—$1.9k
Image gen batch4× GPU, 8 h—$84

Based on current best on-demand prices. Spot pricing typically 40–70% lower. Actual cost depends on provider, region, and availability.

When to choose each GPU

Choose NVIDIA GB300 SXM6 288GB for:

  • Frontier AI training
  • Inference at scale
  • Memory-bound LLM serving

Weaknesses / Limitations

  • Very limited cloud availability — still ramping supply
  • Extremely high cost — enterprise/hyperscaler pricing only
  • Requires specialized NVLink-C2C infrastructure

Choose NVIDIA H200 141GB for:

  • Very large LLMs
  • Memory-bound inference
  • Multi-modal models

Weaknesses / Limitations

  • Very limited cloud availability — fewer providers than H100
  • Premium pricing over H100 with same FP16 TFLOPS
  • 700W TDP — same power envelope as H100 with no efficiency gain