GeForce RTX 5090 specs
NVIDIA · Shipping now · first available Jan 2025. Dense compute, memory, bandwidth and power from the vendor's documents; each precision stays separate.
All accelerators · Compare with RTX PRO 6000 Blackwell Server
Specifications
| Dense BF16 · FP32 accumulate | 209.5 TFLOPS |
|---|---|
| Dense FP8 · FP32 accumulate | 419 TFLOPS |
| Dense FP4 · FP32 accumulate | 1,676 TFLOPS |
| Memory | 32 GB · GDDR |
| Memory bandwidth | 1.79 TB/s |
| Maximum chip power | 575 W |
| Interconnect | PCIe 5.0 |
| Process node | TSMC 4N |
| First availability | Jan 2025 |
| Availability | Shippingest — estimated / secondary-source figure |
| Est. purchase price | ~$1,999est — estimated / secondary-source figureLaunch list price (MSRP) |
Consumer Blackwell with FP4 support. Tensor rates are NVIDIA's FP32-accumulate figures; the card's 2× FP16-accumulate rates are not comparable with data-center specifications.
Dense tensor rates with FP32 accumulation, never sparse or FP16-accumulate headline figures. Memory is in decimal GB. The est marker flags a value its vendor-document review did not confirm.
Rent it
$1.09/GPU-hrServerless / per-second · Beam · observed
No on-demand quote is tracked; this price is a different billing model and is not compared with on-demand rates.
RTX 5090 rental prices →Value ratios
| $ per dense BF16 PFLOP-hour | No on-demand quote |
|---|---|
| Dense BF16 TFLOPS per watt | 0.36 |
| Bandwidth bytes per 1,000 BF16 FLOP | 8.54 |
| $ per dense FP8 PFLOP-hour | No on-demand quote |
| Dense FP8 TFLOPS per watt | 0.73 |
| Bandwidth bytes per 1,000 FP8 FLOP | 4.27 |
| $ per GB of memory per hour | No on-demand quote |
Prices divide the cheapest on-demand list price; theoretical ratios do not measure application speed.
Sources
Specifications reviewed . Dense FP32-accumulate tensor rates: BF16 209.5 TFLOPS, FP8 419 TFLOPS, FP4 1,676 TFLOPS; process TSMC 4N. The whitepaper's 2× FP16-accumulate rates are not comparable with data-center specifications.
- NVIDIA RTX Blackwell GPU Architecture whitepaper V1.1, Appendix A, Table 3 ↗
- NVIDIA GeForce RTX 5090 ↗
First availability: NVIDIA: GeForce RTX 5090 and 5080 out now ↗ (2025-01-30): “out now!”
Purchase price: NVIDIA GeForce RTX 50 Series launch ↗ (2025-01-06): “available on Jan. 30 at $1,999”
Rental price: Beam published pricing ↗ (RTX 5090 PCIe 32 GB serverless GPU task; one GPU; rate published per second, billed by the millisecond).