H100 80GB SXM5 vs RTX 6000 Ada 48GB for Local LLMs
The H100 80GB SXM5 has about 249% more memory bandwidth (3,350 vs 960 GB/s). Batch-1 decode tracks bandwidth closely, so a model that fits both cards should decode roughly 249% faster on the H100 80GB SXM5. The H100 80GB SXM5 fits more of the catalog at Q4/8K — 46 vs 39 of 61 models (80 vs 48 GB VRAM). Largest model that fits the H100 80GB SXM5 but not the RTX 6000 Ada 48GB: gpt-oss-120b (120.4B). MSRP is $30,000 for the H100 80GB SXM5 vs $6,800 for the RTX 6000 Ada 48GB. MSRP is a launch list price, not live retail — Amazon prices move; click through for the current price. Not retail-shoppable: H100 80GB SXM5 (datacenter-class; cloud rental is the realistic way to use it). In the Q4/8K head-to-head below (50 shared models), the H100 80GB SXM5 posts the higher tok/s in 50 rows vs 0. Methodology.