Runpod Our sponsor
USA 1 config 1x
On-Demand
from $0.18
Nvidia RTX 4000 SFF Ada
Low-power professional card for small-form-factor workstations and dense inference nodes.
Latest update just now
Aggregating historical prices...
Key Specifications
RTX 4000 SFF Ada Pricing and Availability
With a spread of only 12%, price optimization yields diminishing returns for the RTX 4000 SFF Ada. The market is relatively flat, with the lowest rate sitting at $0.44/hr per GPU (on-demand).
Heads up: We do our best to keep these specs & prices accurate. However, cloud costs may fluctuate based on region, usage, and other factors not listed here. These are estimates based on common setups and are for informational purposes only. Always verify current rates & exact specs with the provider before provisioning.
Frequently Asked Questions
Why choose the RTX 4000 SFF Ada?
20GB GDDR6 with ECC inside a 70W board power envelope, so it runs in small-form-factor workstations and dense chassis that cannot supply or cool a full-height card. Ada Lovelace FP8 Tensor Cores cover small-model inference, AI development, rendering and viewport work.
When is the RTX 4000 SFF Ada not a good fit?
The 70W envelope costs throughput: single-precision performance is roughly 30% below the full-height RTX 4000 Ada, which shares the same 20GB and pin count. There is no NVLink to pool memory across cards. For continuous cloud inference at a similar power draw, the L4 offers 24GB and higher memory bandwidth.
What size AI models can the RTX 4000 SFF Ada run?
With 20GB of VRAM, the RTX 4000 SFF Ada can typically run 7B to 13B models in FP16, or larger models in 4-bit quantized form.
How much VRAM does the RTX 4000 SFF Ada have?
The RTX 4000 SFF Ada has 20GB of VRAM. Multi-GPU setups increase total memory, but that memory is not automatically pooled across GPUs.
What is the RTX 4000 SFF Ada's memory bandwidth?
The RTX 4000 SFF Ada has 280 GB/s of memory bandwidth. Higher bandwidth helps with faster data transfer between GPU memory and compute cores.
What data types does the RTX 4000 SFF Ada support?
The RTX 4000 SFF Ada supports 7 precision formats. Training: BF16, FP16, TF32, FP32. Inference: FP8, INT4, INT8.
Does the RTX 4000 SFF Ada support NVLink?
No. The RTX 4000 SFF Ada is a PCIe-only GPU with no NVLink, so it is better suited to single-GPU inference and smaller-scale workloads than large distributed training jobs.
How much does the RTX 4000 SFF Ada cost per hour?
RTX 4000 SFF Ada pricing currently ranges from $0.44/hr to $0.50/hr per GPU, depending on the provider, instance type, and billing model.
How much does the RTX 4000 SFF Ada cost per month?
At 720 hours per month, one RTX 4000 SFF Ada can cost between $316.80 to $360.00 per month, depending on the provider. Reserved and spot pricing can lower that further.
Which cloud providers offer the RTX 4000 SFF Ada?
The RTX 4000 SFF Ada is available from 3 cloud providers: Hetzner, Koyeb, Runpod.
Can I rent the RTX 4000 SFF Ada in the cloud?
Yes. We currently track 3 RTX 4000 SFF Ada listings across 3 cloud providers:
Billing type Listings Avg $/GPU/hr On-demand 3 $0.37/hr
Technical Specifications
| GPU Memory | 20GB GDDR6 |
| Memory Interface | 160-bit |
| Memory Bandwidth | 280 GB/s |
| Error Correcting Code | Yes |
| Architecture | NVIDIA Ada Lovelace |
| CUDA Cores | 6,144 |
| Tensor Cores (Fourth-generation) | 192 |
| RT Cores (Third-generation) | 48 |
| Single-Precision Performance | 19.2 TFLOPS |
| RT Core Performance | 44.3 TFLOPS |
| Tensor Performance | 306.8 TFLOPS |
| System Interface | PCIe 4.0 x16 |
| Power Consumption | Total board power: 70W |
| Thermal Solution | Active |
| Form Factor | 2.7" H x 6.6" L, dual slot |
| Display Connectors | 4x mini DisplayPort 1.4a |
| Encode/Decode Engines | 2x encode, 2x decode (+AV1 encode and decode) |
| VR Ready | Yes |
| Graphics APIs | DirectX 12 Ultimate, Shader Model 6.7, OpenGL 4.6, Vulkan 1.3 |
| Compute APIs | CUDA 12.0, OpenCL 3.0, DirectCompute |
| NVIDIA NVLink® | No |
Alternatives to Nvidia RTX 4000 SFF Ada
Nvidia RTX 4000 Ada
Full-height Ada professional card with the same memory capacity at a higher power envelope. From $0.08/hr per GPU across 4 providers.
Nvidia L4
Ada Lovelace cloud inference GPU in a comparable power envelope. From $0.11/hr per GPU across 16 providers.
Nvidia RTX 2000 Ada
Smaller Ada professional card at the same power envelope.. 1 provider, pricing on request.
Nvidia A4000
Previous-generation compact professional GPU. From $0.03/hr per GPU across 10 providers.
Last updated