RTX A4000 vs RTX 6000 Ada
RTX A4000 starts at $0.10/GPU-hour on-demand; RTX 6000 Ada starts at $0.75/GPU-hour on-demand. Compare the exact variants, memory and published specifications below. Peak compute figures do not establish application speed.
RTX A4000
RTX 6000 Ada
Spec sheet
| RTX A4000 | RTX 6000 Ada | |
|---|---|---|
| Spec variant | RTX A4000 | RTX 6000 Ada |
| Architecture | Ampere | Ada Lovelace |
| VRAM (spec variant) | 16 GB | 48 GB |
| FP16 TFLOPS (dense) | 77 | 364 |
| Memory bandwidth | 448 GB/s | 960 GB/s |
| TDP | 140 W | 300 W |
| Interconnect | PCIe Gen4 | PCIe Gen4 |
| Launched | 2021 | 2022 |
Rental price per manufacturer-rated FP16 compute: RTX A4000 $1.30 per PFLOP-hour vs RTX 6000 Ada $2.06. Formula: cheapest $/hr ÷ dense FP16 PFLOPS (1 PFLOP = 1000 TFLOPS). This uses RTX A4000 and RTX 6000 Ada, whose datasheets may differ from the variants above. It is not measured job throughput. Specs: RTX A4000 manufacturer, RTX 6000 Ada manufacturer.
What they cost to rent (2026-10-02)
| RTX A4000 | RTX 6000 Ada | |
|---|---|---|
| Cheapest on-demand $/hr | $0.10 | $0.75 |
| Cheapest spot/community $/hr | $0.17 | $0.74 |
| Cheapest source | TensorDock | TensorDock |
| Providers offering it | 3 | 6 |
Lowest recorded on-demand rate per configuration; highlighted = cheaper. Spot/marketplace tiers shown on the individual pages.
FAQ
RTX A4000 is cheaper: the RTX A4000 rents from $0.10 per GPU-hour and the RTX 6000 Ada from $0.75 on-demand, updated 2026-10-02.
The largest recorded RTX A4000 variant has 16 GB and the largest RTX 6000 Ada variant has 48 GB. Other variants may have less memory. The RTX 6000 Ada has more.
Compare more rental options: the screener · price movers · methodology.