All GPUs
NVIDIA · Blackwell
NVIDIA B200
Blackwell-generation throughput for large-model training and high-volume inference with FP8 and FP4.
- On-demand
- 2 years
- 3 years
Specifications
- Architecture
- Blackwell
- GPU memory
- 180 GB HBM3e
- Memory bandwidth
- 8 TB/s
- GPU-to-GPU
- NVLink 5 · 1.8 TB/s
- Form factor
- SXM (HGX)
Best for
- Large-model training and fine-tuning
- High-throughput LLM inference
- Mixture-of-experts serving