Run DeepSeek-Coder V2 Lite

Ready in about 90 seconds. No install, no setup. From $0.45 per hour.

SOC 2 controls 142 verified providers 99.2% uptime Ready in about 90 seconds

What it does

DeepSeek-Coder V2 Lite is a 16 billion parameter model that only activates a small part of itself for any given request, which makes it much faster than its size suggests.

That speed makes it the right choice for editor autocomplete, where a two second delay is unusable and a two hundred millisecond one is invisible.

It is weaker than the 32 billion parameter models on hard problems. Many people run this for completion and switch to Qwen2.5-Coder for the difficult questions.

What you can ask it

You give it

Inline completion as you type in your editor.

You get back

Suggestions fast enough that you do not wait for them.

You give it

Write a function from this docstring.

You get back

A working implementation plus its type hints.

You give it

Why does this throw on an empty list?

You get back

Points to the unguarded index and shows the fix.

Your setup

WE RUN THIS ON RTX 4090

We run this on an RTX 4090. It writes at about 90 tokens per second, which is fast enough for live autocomplete. $0.45 per hour, billed by the minute.

Why this card?

This model only uses 2.4 billion of its 16 billion parameters per token, but all 16 billion have to be in memory in case they are needed. That is about 11GB compressed, and a 24GB card leaves enough left over for a useful amount of code context. The card is chosen for latency rather than throughput, because for autocomplete the wait on one short answer is the only thing that matters.

Memory
24GB
Speed
about 2 minutes per thousand lines
Rate
$0.45/hr

What will it cost you?

40 thousand lines written 60 minutes $0.45

Billed by the minute on a RTX 4090 at $0.45 per hour. A batch job shuts the GPU off when the last thousand lines is done, so this is the whole cost.

Two ways to run it

OPEN THE APP

Click and use it

Connect your editor to the session and code with completion running.

Get early access Best for exploring.
RUN A BATCH JOB

Hand us the whole pile

Bulk docstring or comment generation across a repository.

Join the list Best for volume. The GPU shuts off automatically when it is done.
587 DeepSeek-Coder V2 Lite jobs run on GPUVault in the last 30 days
Model facts
Parameters16B total, 2.4B active
LicenseDeepSeek License (commercial use permitted)
Memory required11GB minimum, 24GB recommended
Base modelTrained from scratch by DeepSeek
PublisherDeepSeek
Recommended hardwareRTX 4090, 24GB
Hugging Facedeepseek-ai/DeepSeek-Coder-V2-Lite-Instruct
COMING SOON

Want DeepSeek-Coder V2 Lite the day it opens?

We are testing the rental flow with a small group first. Leave an email and we will tell you when DeepSeek-Coder V2 Lite is ready to run. One message, no newsletter.

DeepSeek-Coder V2 Lite questions

Do I need to install anything to use DeepSeek-Coder V2 Lite?

No. We start a machine with DeepSeek-Coder V2 Lite 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 40 thousand lines written?

About $0.45. DeepSeek-Coder V2 Lite takes about 2 minutes per thousand lines on the RTX 4090 we recommend, so 40 thousand lines written is roughly 60 minutes of GPU time at $0.45 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-Coder V2 Lite on a cheaper card?

Sometimes, and the calculator will not always make it look worth it. This model only uses 2.4 billion of its 16 billion parameters per token, but all 16 billion have to be in memory in case they are needed. That is about 11GB compressed, and a 24GB card leaves enough left over for a useful amount of code context. The card is chosen for latency rather than throughput, because for autocomplete the wait on one short answer is the only thing that matters. If you want to try a different card anyway, the advanced catalog lets you pick one and shows the estimated time before you commit.

What are the license restrictions on DeepSeek-Coder V2 Lite?

DeepSeek-Coder V2 Lite is released under the DeepSeek License (commercial use permitted). That is not a permissive license, so read it before you use the output in paid work. We show the license on every model page precisely because this catches people out, and there is usually a permissively licensed alternative in the same category.

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.

Own a GPU that sits idle? Put it to work. The average provider earns $180 to $420 per month per card.

List your GPU