Skip to content

Hopper H100 SXM5 vs Ampere A100 SXM4

Complete side-by-side comparison of specs, performance, memory, power efficiency, and pricing.

NVIDIA

Hopper H100 SXM5

73

Spec Wins

NVIDIA

Ampere A100 SXM4

61

Detailed Specifications

SpecHopper H100 SXM5Ampere A100 SXM4
ArchitectureHopper Ampere
Memory80GB HBM3 80GB HBM2e
Memory Bandwidth3,350 GB/s 2,039 GB/s
FP16 TFLOPS1,979 312
FP8 TFLOPS3,958 0
BF16 TFLOPS1,979 624
INT8 TOPS3,958 1,248
TDP700W 400W
InterconnectNVLink 4.0 (900 GB/s) (900 GB/s) NVLink 3.0 (600 GB/s) (600 GB/s)
Perf Score73 61
EcosystemCUDA CUDA
Est. Price$25,000 $12,000

Hopper H100 SXM5 — Best For

LLM TrainingHPC

Ampere A100 SXM4 — Best For

TrainingFine-tuning

Who Should Choose Each GPU?

Choose Hopper H100 SXM5 if you…

  • Need maximum CUDA/TensorRT/vLLM ecosystem compatibility
  • Prioritize raw FP8 throughput (3,958 vs 0 TFLOPS)
  • Running LLM Training workloads
  • Running HPC workloads

Choose Ampere A100 SXM4 if you…

  • Need maximum CUDA/TensorRT/vLLM ecosystem compatibility
  • Have power-constrained data centers (400W vs 700W TDP)
  • Working with a tighter CapEx budget (lower list price)
  • Running Training workloads
  • Running Fine-tuning workloads

Verdict

The Hopper H100 SXM5 and Ampere A100 SXM4 target different priorities. Both GPUs offer the same 80GB VRAM, so the decision comes down to ecosystem and compute throughput. For training throughput, the Hopper H100 SXM5's 3,958 FP8 TFLOPS outpaces the Ampere A100 SXM4's 0 TFLOPS. Both GPUs use CUDA, so ecosystem switching cost is not a factor. Use our TCO Calculator to model the full 3-year cost difference for your specific utilization and power costs.

Hopper H100 SXM5 vs Ampere A100 SXM4: Common Questions

Which is faster, Hopper H100 SXM5 or Ampere A100 SXM4?+

In FP8 throughput, the Hopper H100 SXM5 leads with 3,958 TFLOPS vs 0 TFLOPS. For LLM inference, memory capacity and bandwidth often matter more than raw TFLOPS — the Ampere A100 SXM4 has more VRAM (80GB).

Is Hopper H100 SXM5 or Ampere A100 SXM4 better for LLM training?+

For LLM training at scale, the Hopper H100 SXM5 has higher raw throughput. However, the choice also depends on your software stack: Hopper H100 SXM5 offers CUDA compatibility with the widest framework support (PyTorch, JAX, TensorRT).

What is the price difference between Hopper H100 SXM5 and Ampere A100 SXM4?+

The Hopper H100 SXM5 is estimated at $25,000 per unit and the Ampere A100 SXM4 at $12,000. Actual pricing varies by vendor, volume, and configuration. Check our Buy page for current reseller pricing.

Which GPU is more power efficient, Hopper H100 SXM5 or Ampere A100 SXM4?+

The Ampere A100 SXM4 has a lower TDP (400W vs 700W). Performance-per-watt depends on your workload — for FP8 inference, divide TFLOPS by TDP: Hopper H100 SXM5 = 5.7 TFLOPS/W vs Ampere A100 SXM4 = 0.0 TFLOPS/W.

Ask AI Advisor