
B200
Next-generation data center GPU based on Blackwell architecture that features 180 GB of HBM3e memory with 8TB/s bandwidth, delivering up to 20 petaFLOPS of FP4 AI compute performance.
GPU Benchmarks
Compare performance across LLMs and image models to find the best GPU for your workload.
Runpod benchmark measurements from September 2026, using vLLM.

Next-generation data center GPU based on Blackwell architecture that features 180 GB of HBM3e memory with 8TB/s bandwidth, delivering up to 20 petaFLOPS of FP4 AI compute performance.

Compact professional GPU based on Ada Lovelace architecture with 16GB GDDR6 memory and 2,816 CUDA cores for AI workloads, machine learning, and professional applications in small form factor systems.

High-efficiency LLM processing at 90.98 tok/s.
Runpod benchmark measurements from September 2026, using Hugging Face Diffusers.

Unmatched image gen speed with 49.9 images per minute.

AI image processing at 40.3 images per minute.

Pro-grade performance with 36 images per minute.
Case Studies