NVIDIA DGX H200 (8x H200, 1,128 GB) vs NVIDIA DGX Station GB300 (748 GB)

NVIDIA DGX H200 (8x H200, 1,128 GB) · NVIDIA DGX Station GB300 (748 GB)

Verdict
  • throughputThe NVIDIA DGX H200 (8x H200, 1,128 GB) delivers approximately 809% higher aggregate decode throughput across the 8 models both machines can run (geometric mean of the per-model ratios).
  • capacityBoth machines run all 8 compared models, so this is a pure speed and cost decision rather than a capability one.
  • costAt 8 hours a day, the NVIDIA DGX Station GB300 (748 GB) has approximately 75% lower total monthly cost ($1,182 against $4,743), for a C corporation in United States (federal).
  • powerThe NVIDIA DGX H200 (8x H200, 1,128 GB) draws 8500 W under load against 1350 W for the NVIDIA DGX Station GB300 (748 GB) — a 6.3× difference that shows up in the electricity line and in how loud and hot the room gets.
  • evidenceEvery throughput figure in this comparison is estimated. Treat the ranking as indicative and the margins as unreliable.
  • recommendationThere is no single answer: the NVIDIA DGX H200 (8x H200, 1,128 GB) is faster, the NVIDIA DGX Station GB300 (748 GB) is cheaper to own, and the NVIDIA DGX H200 (8x H200, 1,128 GB) gives more throughput per unit of purchase currency. Pick on whichever constraint actually binds you — interactive latency, monthly budget, or capacity headroom for larger models later.
Every statement above is generated from the calculations on this page by fixed rules — no prose is written for individual comparisons. If the numbers change, the verdict changes with them.
Hardware
NVIDIA DGX H200 (8x H200, 1,128 GB)NVIDIA DGX Station GB300 (748 GB)
Purchase price estimate$359,430 *$90,082 *
Memory1,128 GB748 GB
Memory bandwidth38,400 GB/s7,100 GB/s
Achieved bandwidth29,952 GB/s (78%)1,065 GB/s (15%)
Nominal power10200 W1600 W
Measured load power8500 W1350 W
Idle power1600 W180 W
Architecture8× GPUunified memory
Model performance8 models run on both
Kimi K2.6GLM-5.3MiniMax M3GLM-5.3 FlashDeepSeek-V4.1 FlashDeepSeek-V3.2DeepSeek-V4 FlashHunyuan Hy3
ModelNVIDIA DGX H200 (8x H200, 1,128 GB)NVIDIA DGX Station GB300 (748 GB)Winner
Kimi K2.6
1.0T (32B active)
~367–528 t/s
Q4_K_M
~36.6–52.6 t/s
Q4_K_M
NVIDIA DGX H200 (8x H200, 1,128 GB) +904%
GLM-5.3
753.33B (40B active)
~254–365 t/s
Q8_0
~29.5–42.5 t/s
Q4_K_M
NVIDIA DGX H200 (8x H200, 1,128 GB) +760%
MiniMax M3
427.04B (23B active)
~334–481 t/s
Q8_0
~29.3–42.1 t/s
Q8_0
NVIDIA DGX H200 (8x H200, 1,128 GB) +1042%
GLM-5.3 Flash
321.323B (18B active)
~369–531 t/s
Q8_0
~37.1–53.3 t/s
Q8_0
NVIDIA DGX H200 (8x H200, 1,128 GB) +895%
DeepSeek-V4.1 Flash
763.205B (16B active)
~451–649 t/s
Q4_K_M
~70–101 t/s
Q4_K_M
NVIDIA DGX H200 (8x H200, 1,128 GB) +545%
DeepSeek-V3.2
685.397B (37B active)
~265–382 t/s
Q8_0
~31.8–45.8 t/s
Q4_K_M
NVIDIA DGX H200 (8x H200, 1,128 GB) +734%
DeepSeek-V4 Flash
304.18B (13B active)
~411–592 t/s
Q8_0
~50.4–72.6 t/s
Q8_0
NVIDIA DGX H200 (8x H200, 1,128 GB) +715%
Hunyuan Hy3
298.786B (21B active)
~347–500 t/s
Q8_0
~32–46 t/s
Q8_0
NVIDIA DGX H200 (8x H200, 1,128 GB) +987%
Differences under 10% are reported as a tie: they are smaller than the spread between published benchmarks of the same configuration, so calling a winner would be false precision. Rows where a model fits on only one machine are decided by capacity, not speed.
Economics — United States (federal), C corporation, 8h/day
NVIDIA DGX H200 (8x H200, 1,128 GB)NVIDIA DGX Station GB300 (748 GB)
Purchase price estimate$359,430$90,082
Reclaimable VAT$0$0
Net cash at purchase$359,430$90,082
Monthly depreciation$5,271.64$1,321.20
Monthly electricity$73.43 (524 kWh)$10.20 (73 kWh)
Tax programme§179 100.0%§179 100.0%
Tax-programme benefit$75,480$18,917
Monthly economic cost$4,742.58$1,182.12
Total over 5 years$284,555$70,927
Tax assumptions last verified 2026-09-07. Federal planning estimate only, not tax advice. State and local income tax, sales/use tax and incentives are excluded. It assumes 100% business use, a Section 179 election, enough business income to use it, and that the full annual limit remains available.