Nvidia B100

Nvidia B100

Blackwell data center GPU for next-generation AI training and inference.

Arch
Blackwell
Memory
192 GB HBM3e
Bandwidth
8.00 TB/s
Released
Q4 2024

At a glance

We don't currently track any B100 offers. See alternative GPUs.

Technical Specifications

Nvidia B100 · Per GPU

Compute · dense
FP4 7,000 TFLOPS 14,000 with sparsity
FP8 3,500 TFLOPS 7,000 with sparsity
FP16 / BF16 1,750 TFLOPS 3,500 with sparsity
INT8 3,500 TOPS 7,000 with sparsity
FP32 60 TFLOPS
FP64 30 TFLOPS
Precision support FP4FP6FP8FP16BF16TF32FP32FP64INT8
Memory
Capacity 192 GB HBM3e
Bandwidth 8,000 GB/s
ECC Yes
Silicon
Architecture Blackwell
Process TSMC 4NP
Transistors 208 billion
Fabric and host
GPU interconnect NVLink 1,800 GB/s
Host interface PCIe 5.0 x16
Power
Cooling Passive
Platform
Partitioning MIG, up to 7 instances

Source: official Nvidia B100 datasheet.

Frequently Asked Questions

Why choose the B100?

192GB HBM3e with Blackwell architecture FP4 Tensor Cores. 1800 GB/s NVLink for massive multi-GPU scaling. Significant inference throughput improvement over H100 with FP4 precision.

When is the B100 not a good fit?

New generation with limited provider availability. High cost per hour. For inference on models under 70B or workloads that don't benefit from FP4, the H100 or L40S is more cost-effective.

Alternatives to Nvidia B100

Last updated