NVIDIA B200 SXM
NVIDIA's flagship Blackwell-architecture datacenter GPU with 192 GB of HBM3e memory and 8 TB/s of memory bandwidth. Designed for trillion-parameter LLM training and large-scale inference, with 5th-generation Tensor Cores and FP4 precision support.
VRAM
192 GB
Memory
HBM3e
Bandwidth
8,000 GB/s
FP16
4,500 TF
TDP
1000 W
Memory and compute
Architecture and power
Notes
Compute throughput shown with 2:4 structured sparsity. Two-die package connected via NV-HBI. Liquid-cooled HGX SXM module.
Models that may fit on NVIDIA B200 SXM
Estimates based on INT8 quantization. Actual fit depends on framework and batch size.
More GPUs by NVIDIA
The manufacturer's other cards in the catalog, largest memory first, with bandwidth and power where published.
Frequently asked questions about NVIDIA B200 SXM
Answered from the entry's own figures: memory, which models fit, bandwidth, power and the card's class.
NVIDIA B200 SXM has 192 GB of HBM3e at 8,000 GB/s. At INT8 a model's weights take about one byte per parameter, so the card holds roughly a 153B-parameter model with room for context; larger models need several cards or a lower precision.
From the catalog's estimates at INT8, Gemma 4 31B IT, Llama Nemotron Embed VL 1B V2, GLM Z1 9B 0414 and NVIDIA Nemotron Nano 12B V2 VL BF16 fit on a single NVIDIA B200 SXM with their default context. The list on this page shows the estimated memory each takes; the capacity planner sizes any model against this card for your own context length, batch size and request rate.
8,000 GB/s. Bandwidth bounds how fast a model generates tokens, because every token reads the whole set of weights and the KV cache from memory; at this rate a 153B model at INT8 could be read on the order of 52 times a second on one card, before batching and framework efficiency.
NVIDIA B200 SXM has a TDP of 1000 W, with a maximum of 1200 W, in the SXM form factor. Budget the TDP per card plus host overhead when sizing a server's power and cooling.
NVIDIA B200 SXM is a datacenter accelerator, meant for servers and multi-GPU serving, built on the Blackwell architecture and released in 2025. It supports a multi-GPU interconnect at 1,800 GB/s, so a model larger than one card can be split across several.
Specifications from the manufacturer's published material; a figure that is not published is shown as such, and a derived one is explained in the notes.
Added Apr 30, 2026 · Last updated Apr 30, 2026
Explore more
Companion tools that draw on the same catalog and routing engine.
Put intelligence in the middle.
One gateway in front of every model, with your policies applied and every decision on record. Start with $5 of credit and 5,000 routing decisions a month, no card required.