> For the complete documentation index, see [llms.txt](https://docs.quantaria.xyz/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.quantaria.xyz/models.md).

# Models

Quantaria has five model lines, named for the physics they run on. Hilbert-1 Nano and Base are live on the [Quantaria Console](/quantaria-console.md). The other lines are planned or in research.

| Line                | Role                                                                     | Status                            |
| ------------------- | ------------------------------------------------------------------------ | --------------------------------- |
| [Hilbert](#hilbert) | Flagship: unified text and image generation from one set of tensor cores | Nano and Base live; Large planned |
| [Schmidt](#schmidt) | Compact text model for laptops and phones                                | Planned                           |
| [Dirac](#dirac)     | Hybrid quantum-classical layers                                          | Research                          |
| [Born](#born)       | Born-machine generative heads                                            | Research                          |
| [Bond](#bond)       | Open-source PyTorch library behind every model                           | Planned                           |

## Hilbert

The flagship line. One shared tensor-network backbone generates both text and images, trained with a [Transfusion-style objective](https://github.com/bhavy2908/dataai_docs/tree/main/architecture/multimodal-backbone.md).

|                        | Hilbert-1 Nano           | Hilbert-1 Base           | Hilbert-1 Large          |
| ---------------------- | ------------------------ | ------------------------ | ------------------------ |
| Parameters             | \~100M                   | \~1B                     | Set after Base results   |
| Text context           | 2K tokens                | 8K tokens                | TBD                      |
| Image output           | 32×32                    | 256×256                  | TBD                      |
| Weights                | MPS/MPO                  | MPS/MPO                  | MPS/MPO                  |
| Attention              | Tensor Product Attention | Tensor Product Attention | Tensor Product Attention |
| Objective              | Transfusion-style        | Transfusion-style        | Transfusion-style        |
| Rank-adaptive χ        | Yes                      | Yes                      | Yes                      |
| MPO mixture-of-experts | No                       | Yes                      | TBD                      |
| Hybrid quantum layers  | No                       | No                       | Research                 |
| Status                 | Live on console          | Live on console          | Planned                  |

Hilbert-1 Nano is trained on TinyStories text and CIFAR-scale images and is compared against a dense transformer of the same size. Training data for Base and Large has not been published.

{% hint style="info" %}
Benchmarks against same-size dense transformers (loss curves, perplexity, FID and memory use) are published as each model is evaluated, whether the result favours the tensor-network model or not.
{% endhint %}

## Schmidt

A compact, text-only model for long context on laptops and phones. Low-χ cores and Tensor Product Attention keep both the weights and the KV cache small. Status: **planned**.

## Dirac

Hybrid layers built as quantum circuit, then MPO, then quantum circuit, with Cayley-parameterized unitary adapters. Development starts on circuit simulators and moves to cloud quantum processors (QPUs) as hardware allows. Status: **research**. See [Quantum Path](https://github.com/bhavy2908/dataai_docs/tree/main/architecture/quantum-path.md).

## Born

Born-machine generative heads that sample images from an MPS with probability |ψ(x)|², the way a measurement samples a quantum state. Status: **research**.

## Bond

The open-source PyTorch library behind every Quantaria model: tensor-train and MPO layers, rank-adaptive training, canonical gauging and fused contraction kernels. Status: **planned** open-source release.

> **TODO (Quantaria team):** add the repository link and license for Bond once it is released.


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://docs.quantaria.xyz/models.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.