> ## Documentation Index
> Fetch the complete documentation index at: https://docs.os3compute.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Overview

> Call DeepSeek V4.1 Flash on OS3's GPUs with one key. What it costs, its limits and what to do when a call fails.

OS3 serves open models on its own GPUs in the United States. You call them with an OpenAI or Anthropic SDK and an OS3 key. You need no provider account and no training setup.

## Three steps to a first answer

**1. Create a key**

In the portal: **Keys → Create key**. It is shown once: keep it in your environment, never in code.

**2. Paste the code**

From the [quickstart](https://docs.os3compute.com/quickstart.md). Your key is the API key, and the model is its id.

**3. Get an answer**

Your first US$10 is on us: a new organisation gets US$10 of credit when OS3 sets it up.

Then add a card or buy credit in **Usage**, before the free credit runs out. See [Pricing](https://docs.os3compute.com/pricing.md).

> [card] [Quickstart](https://docs.os3compute.com/quickstart.md)
>
> curl, Python and TypeScript: chat completions, Messages, Responses, thinking and the fast id.

> [card] [DeepSeek V4.1 Flash 1017](https://docs.os3compute.com/models/deepseek-v4-1-flash-1017.md)
>
> The ids, context, images, tools, thinking and streaming.

> [card] [Pricing](https://docs.os3compute.com/pricing.md)
>
> Prices per token, how you pay, and the plans.

> [card] [Limits by plan](https://docs.os3compute.com/limits.md)
>
> Requests and tokens a minute, and requests in flight.

> [card] [Errors](https://docs.os3compute.com/errors.md)
>
> Each status and code, with what to do.

> [card] [Your data](https://docs.os3compute.com/data.md)
>
> What OS3 keeps, and Processing, which is off until you turn it on.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.