H200 SXM
LLM inference · fine-tuning
$1,832/ 30 days
View GPU H200 SXM vs RTX 6000 Ada
Compare memory, node configurations, and fixed weekly and monthly cost before choosing between NVIDIA H200 SXM and NVIDIA RTX 6000 Ada.
H200 SXM provides 93 GB more VRAM per card. RTX 6000 Ada costs $1,464 less per 30-day term . Choose based on whether memory headroom or lower fixed cost is the binding constraint.
LLM inference · fine-tuning
3D rendering · fine-tuning
| Specification | H200 SXM | RTX 6000 Ada |
|---|---|---|
| Architecture | Hopper | Ada |
| VRAM | 141 GB HBM3e | 48 GB GDDR6 ECC |
| Host resources / card | 24 vCPU · 192 GB RAM | 16 vCPU · 128 GB RAM |
| Local NVMe / card | 3.84 TB | 2 TB |
| 7-day price / card | $512 | $104 |
| 30-day price / card | $1,832 | $368 |
| 4-GPU node / 30 days | $6,742 | $1,354 |
| Catalog quantity | 265 cards | 410 cards |