Skip to main content

NVIDIA H200 SXM 141GB vs NVIDIA B200 — Specs & Performance Compared

Side-by-side comparison of 2 GPU parts: NVIDIA H200 SXM 141GB vs NVIDIA B200. All released in 2024. Specs, synthetic benchmarks, AI inference tok/s, and Amazon availability — click into any column header for the per-SKU benchmark page.

vs

*Price sourced from Amazon.com. Price and availability subject to change.

Specs side-by-side

SpecNVIDIA H200 SXM 141GBNVIDIA B200
vram gb 141192
vram type HBM3eHBM3e
tdp w 7001000
cuda cores 1689616896
tensor cores 528—
l2 cache mb —50
process nm 45

Synthetic benchmarks

Higher is better. Per-row scores come from public benchmark databases (PassMark, TechPowerUp, Geekbench, Cinebench). Empty cells mean we don't yet have that benchmark on file for the SKU.

BenchmarkNVIDIA H200 SXM 141GBNVIDIA B200
3DMark Time Spy 15,200 points—
MLPerf Inference v4.0 — Llama 2 70B Offline Scenario 31,712 tokens/s—
MLPerf Inference v5.0 — Llama 2 70B Server Scenario 33,000 tokens/s—
MLPerf Inference v5.1 — Llama 2 70B Offline —102,725 tok/s
MLPerf Inference v6.0 — DeepSeek R1 Offline —58,582 tok/s
MLPerf Inference v6.0 — gpt-oss 120B Offline —85,921 tok/s
Port Royal (RT) 11,800 points—

AI inference (tokens/sec)

Top community-reported tokens-per-second results for popular LLMs. See the per-SKU benchmark page for the full set including quantization, prompt length, and source citations.

NVIDIA H200 SXM 141GB

  • llama2:70b: 33072.0 tok/s
  • llama2:13b: 11819.0 tok/s
  • llama2:13b: 11819.0 tok/s

All NVIDIA H200 SXM 141GB AI benchmarks →

NVIDIA B200

  • llama3.1:8b: 160403.0 tok/s
  • Mixtral 8x7B: 129047.0 tok/s
  • mixtral:8x7b: 128148.0 tok/s

All NVIDIA B200 AI benchmarks →

Keep exploring

More guides & deep dives from the SpecPicks archive

Browse all articles & guides →

More reviews from the SpecPicks archive

Browse all reviews →

More buying guides from SpecPicks

Browse all buying guides →

Hardware benchmark data on SpecPicks

All benchmarks →