Salad
Our sponsor- On-Demand
- from $0.17
- Spot
- from $0.08
Budget Blackwell GPU for local AI experimentation.
Weekly median price per GPU per hour · Get the data
By provider shows one card per company, placed where its best offer ranked. By configuration lists every offer.
What you can rent comes first: in stock, then waitlist, then not reported, then out of stock. Priced offers rank ahead of quote-only, and on-demand ahead of other billing types.
Within a group, five factors set the order:
We match the provider, its country, GPU model, form factor, billing type, availability and instance name. Matching is partial and case-insensitive. Everyday words work too, so "interruptible" finds spot and "sold out" finds out of stock.
No providers match .
Heads up: A provider's own page may quote a different figure, for example on tax, region or a promotion. A monthly-only plan shows a derived hourly rate. Verify before provisioning. More on how we price.
One RTX 5070 has 12 GB of VRAM. In practice, that's enough memory for roughly 9B parameters at 4-bit or 0.4B at 16-bit, assuming a 32K context. Below are some open-weight LLMs, with the estimated memory and GPUs each one needs.
| Model | Memory (INT4 / FP4) | RTX 5070s needed | Cost /hr | Cost /mo |
|---|---|---|---|---|
|
|
18 GB
|
2
|
–
|
–
|
|
|
22 GB
|
2
|
–
|
–
|
|
|
67 GB
|
7
|
–
|
–
|
|
|
178 GB
|
17
|
–
|
–
|
|
|
239 GB
|
23
|
–
|
–
|
|
|
306 GB
|
29
|
–
|
–
|
|
|
1,544 GB
|
143
|
–
|
–
|
Estimates based on the median on-demand rate. Memory is weights plus FP8 KV cache at 32K context per request (FP16/BF16 in the 16-bit column). GPU counts assume 90% of advertised VRAM is usable. No guarantee of runtime support, usable performance, or that a matching quantized build exists. Pricing methodology.
Nvidia RTX 5070 · Per GPU
| Compute · dense | |
|---|---|
| FP4 | 493.9 TFLOPS 987.8 with sparsity |
| FP8 | 123.5 TFLOPS 246.9 with sparsity |
| FP16 / BF16 | 61.7 TFLOPS 123.5 with sparsity |
| INT8 | 246.9 TOPS 493.9 with sparsity |
| FP32 | 30.9 TFLOPS |
| Precision support | FP4FP8FP16BF16TF32FP32INT8 |
| Memory | |
|---|---|
| Capacity | 12 GB GDDR7 |
| Bandwidth | 672 GB/s |
| Bus width | 192-bit |
| ECC | No |
| Silicon | |
|---|---|
| Architecture | Blackwell |
| Process | TSMC 4N |
| Transistors | 31.1 billion |
| Shader cores | 6,144 CUDA cores |
| Matrix cores | 192 Tensor cores |
| Compute units | 48 SMs |
| Fabric and host | |
|---|---|
| Host interface | PCIe 5.0 |
| Power | |
|---|---|
| Board power | 250 W |
| Cooling | Active |
Source: official Nvidia RTX 5070 datasheet.
As of October 3, 2026, the median on-demand price is $0.20 per GPU per hour across 3 providers.
| Billing type | Configs | Median /GPU/hr | Cheapest |
|---|---|---|---|
On-demand | 8 | $0.20 | $0.17 |
Spot | 6 | $0.08 |
No median for spot: we only show this when at least 3 providers list the GPU on that billing type.
At 720 hours per month, one RTX 5070 costs an estimated $144 at the median on-demand price. Cheapest verified in stock: $122 per month on-demand, $65 spot.
GetDeploying currently tracks RTX 5070 configs from 3 providers. Vast.ai and Theta EdgeCloud have verified in-stock on-demand configs. See the full price comparison above for every provider and config.
One RTX 5070 runs models up to roughly 9B parameters at 4-bit quantization or 0.4B at 16-bit, assuming a 32K context. Larger models run across multiple GPUs: the model table above shows the estimated memory and GPU count for popular open-weight LLMs.
12GB GDDR7 with Blackwell FP4 Tensor Cores. Budget entry point for local AI experimentation with latest-gen architecture.
Only 12GB VRAM severely limits AI model sizes. Subject to Nvidia's consumer EULA, which may restrict data center use. Better suited for local experimentation.
Last updated