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

# Models

> Learn where to browse models in NagaAI, how free and paid access works, and how to choose a model for your task.

<Tip>
  If you want the fastest way to understand what is available on NagaAI,
  start with [naga.ac/models](https://naga.ac/models). It is the main product
  catalog for comparing providers, pricing, capabilities, and exact model IDs
  before you put anything into code.
</Tip>

## Where to look

For most people, the models page is enough.

Use it to:

* browse the full catalog
* compare providers and pricing
* see which models support tools, multimodal input, reasoning, or other capabilities
* find free models
* copy the exact model ID you want to use

## Free and paid models

NagaAI has both free and paid models.

Free models are marked with the `:free` suffix. They are good for testing, experimenting, and lighter workloads.

Paid models require a positive account balance. They give you access to the broader catalog and are the usual choice for production work.

If your goal is simply "find a free model I can try right now", the quickest path is to open [naga.ac/models](https://naga.ac/models) and filter for `:free`.

## How to think about model choice

The first question usually is not "which provider do I want?" It is "what am I building?"

If you are building chat, assistants, or agent-style workflows, you will usually be looking at models you can use with [Responses API](/api/responses).

If you are building retrieval or RAG, you need an embeddings model.

If you are working with transcription, translation, or text-to-speech, you need an audio model.

If you need safety checks, you need a moderation model.

If you want to generate or edit images, you need an image model.

<CardGroup cols={2}>
  <Card title="LLM Apps" icon="bolt" href="/api/responses">
    Use `Responses API` for chat, assistants, tools, and multimodal workflows.
  </Card>

  <Card title="Embeddings" icon="share-nodes" href="/api/embeddings">
    Use an embeddings model for retrieval, ranking, semantic search, and RAG.
  </Card>

  <Card title="Audio" icon="microphone" href="/api/audio">
    Use audio models for speech-to-text, translation, and text-to-speech.
  </Card>

  <Card title="Moderation" icon="shield" href="/api/moderations">
    Use moderation models when you need text or image safety checks.
  </Card>

  <Card title="Images" icon="image" href="/api/images">
    Use image models when you need image generation or image edits.
  </Card>
</CardGroup>

So the practical flow is simple: first decide what kind of task you have, then use the models page to compare the models that fit that task.

## What matters when choosing

When you compare models on NagaAI, the things that usually matter are:

* whether the model is free or paid
* whether it supports the capability you need
* how much it costs
* whether it is a good fit for experimentation or production

For production, it is usually better to pin the exact model ID in your code instead of relying on a loose name.

## If you need the catalog in code

The website is the best place to explore models as a human.

`/v1/models` exists for the cases where you want the live catalog programmatically, for example in scripts, internal dashboards, or dynamic model selectors.

<CodeGroup>
  ```python Python theme={null}
  import requests

  response = requests.get("https://api.naga.ac/v1/models")
  response.raise_for_status()

  models = response.json()
  print(models["data"][0]["id"])
  ```

  ```javascript Node.js theme={null}
  const response = await fetch('https://api.naga.ac/v1/models');

  if (!response.ok) {
    throw new Error(`Request failed: ${response.status}`);
  }

  const models = await response.json();
  console.log(models.data[0].id);
  ```

  ```bash cURL theme={null}
  curl https://api.naga.ac/v1/models
  ```
</CodeGroup>

The endpoint can be called anonymously. Depending on the account context, the returned catalog may be limited to free variants.

## Related docs

* [Choose an API](/get-started/choose-an-api)
* [Pricing and Billing](/build/billing)
* [Rate Limits](/build/rate-limits)
* [Responses API](/api/responses)
* [Embeddings API](/api/embeddings)
* [List models reference](/api-reference/endpoints/models/list)