# Run DeepSeek-R1 Distill 70B

Thinks through a problem before answering. Slower, and better on hard questions. Ready in about 90 seconds on an A100 80GB, from $1.35/hr. No install, no code.

## What it does

DeepSeek-R1 Distill 70B works through a problem step by step before it gives you an answer, and it shows you that working.

On hard questions with a single right answer, that makes it more reliable than a model of the same size that answers immediately.

It is slower and more expensive per answer, because the thinking is real generated text that you pay for. Use it for the difficult 5% of your work, not the routine 95%.

## Your setup

We run DeepSeek-R1 Distill 70B on an A100 80GB with 80GB of memory, at $1.35 per hour, billed by the minute. It takes about 55 seconds per question.

The memory need is the same as any 70 billion parameter model, roughly 40GB compressed. The reason it wants the larger card is the length of its answers. Reasoning traces run long, and every word generated adds to the context that has to stay in memory. An 80GB card lets a long chain of reasoning finish rather than getting cut off.

## What it costs

200 hard questions answered takes about 3 hours 3 minutes of GPU time, which is about $4.13.

## Two ways to run it

- **Open the app.** A chat window that shows the reasoning above the final answer.
- **Run a batch job.** Send a list of problems and collect answers with the working attached.

## Model facts

| Field | Value |
|---|---|
| Parameters | 70B |
| License | MIT |
| Memory required | 48GB minimum (4-bit), 80GB recommended |
| Base model | Llama 3.3 70B, distilled from DeepSeek-R1 |
| Publisher | DeepSeek |
| Recommended hardware | A100 80GB, 80GB |
| Category | [Language models](https://gpuvault.io/working-on/language-models/) |

## Questions people ask

### Do I need to install anything to use DeepSeek-R1 Distill 70B?

No. We start a machine with DeepSeek-R1 Distill 70B already loaded and hand you a link. Everything runs in your browser, there is nothing to download, and nothing is left on your computer afterwards. It works the same on a Mac, a Windows laptop, or a Chromebook. A workspace is usually ready in about 90 seconds.

### How much does it cost to run 200 hard questions answered?

About $4.13. DeepSeek-R1 Distill 70B takes about 55 seconds per question on the A100 80GB we recommend, so 200 hard questions answered is roughly 3 hours 3 minutes of GPU time at $1.35 per hour. Billing is by the minute, and a batch job shuts the GPU off the moment the last item finishes. The calculator above works this out for your own numbers.

### Can I run DeepSeek-R1 Distill 70B on a cheaper card?

Sometimes, and the calculator will not always make it look worth it. The memory need is the same as any 70 billion parameter model, roughly 40GB compressed. The reason it wants the larger card is the length of its answers. Reasoning traces run long, and every word generated adds to the context that has to stay in memory. An 80GB card lets a long chain of reasoning finish rather than getting cut off. If you want to try a different card anyway, the advanced catalog lets you pick one and shows the estimated time before you commit.

### Can I use DeepSeek-R1 Distill 70B commercially?

Yes. DeepSeek-R1 Distill 70B is released under the MIT license, which permits commercial use with no conditions worth worrying about. You can use the output in paid client work and in products you sell.

### What happens to my files after I stop?

Anything you save into your workspace folder stays in your account and is there when you come back. Everything else is destroyed: the machine is terminated and its disk is wiped. Batch job inputs and outputs are deleted from our storage 24 hours after you download them. We do not read your files or use them to train anything.

### What if it fails partway through?

You are not charged for work that did not complete, and failed sessions are refunded automatically. Batch jobs write each output as it finishes rather than at the end, so a failure at item 380 of 400 leaves you 379 usable files and a message saying what broke. Restarting picks up where it stopped rather than redoing the work.

---

*Source: https://gpuvault.io/model/deepseek-r1-distill-70b/ — GPUVault, GPU rental by the minute.*