NVIDIA GeForce RTX 5090
Consumer
VRAM
32 GB
Memory
GDDR7
Bandwidth
1792 GB/s
TDP
575W
Suitable Workloads
Smaller Language Models
Inference for 7B-13B parameter models
Key Highlights
- Built on Blackwell architecture
- 21,760 compute cores for parallel processing
- PCIe form factor
Memory
VRAM32 GB
Memory TypeGDDR7
Memory Bandwidth1792 GB/s
Compute Performance
FP32 (Single Precision)104.8 TFLOPS
FP16 (Half Precision)419 TFLOPS
BF16419 TFLOPS
Architecture
ArchitectureBlackwell
Compute Cores21,760
Release Year2025
Power & Physical
TDP575 W
Form FactorPCIe
PCIe GenerationPCIe 5.0
Notes
FP16 dense tensor with FP16 accumulate. The highest-VRAM consumer card; common self-host choice for quantized 30B-70B models.
Models That May Fit on NVIDIA GeForce RTX 5090
Estimates based on INT8 quantization. Actual fit depends on framework and batch size.
Explore More
Added Jul 30, 2026
Last updated: Jul 30, 2026
Your AI stack shouldn't stand still.
Every month new models become cheaper, faster, and more capable. Inferbase ensures your application automatically benefits without changing a single API call.