# Mogan AI: lavoir

> A single-pass decision encoder on ModernBERT-large that returns a calibrated probability for every option and a value-of-information score for each candidate clarifying question, so it can ask, decide or hand off. Uses Laya's typed question format.

- Page: https://systemonemodels.ai/moganai/lavoir
- API: https://api.systemonemodels.ai/v1/models/moganai/lavoir
- Download: `pip install systemonemodels && systemone pull moganai/lavoir`

## Facts

| | |
|---|---|
| Maker | Mogan AI (https://systemonemodels.ai/moganai) |
| Decides | choice, score, noul, route |
| Architecture | lavoir |
| Base model | answerdotai/modernbert-large |
| Context | 1K tokens |
| Licence | cc-by-nc-4.0 |
| Availability | Open weights |
| Released | 2026-09-25 |
| Latest version | 2026.09 |

## Reported evaluation

Suite: SGD first user turn, zero-shot (3,812 conversations, 20 services), Mogan AI's evaluation. Numbers are the publisher's own.

- Decision accuracy: 94.2%

- Median latency: 31 ms
- p95 latency: 142 ms

## Model card

<!-- generated by scripts/seed_catalog.py; edit content/models/catalog.yaml -->

# Lavoir

A single-pass decision encoder on ModernBERT-large that returns a calibrated probability for every option and a value-of-information score for each candidate clarifying question, so it can ask, decide or hand off. Uses Laya's typed question format.

Follows Laya's decision-head design, input format and loss, but trains its own heads from ModernBERT-large and uses none of Laya's weights. A small value-of-information head estimates how much asking each missing slot would raise the probability of the right decision. Described in arXiv 2609.30706. Zero-shot on SGD, Mogan AI reports .942 accuracy while asking in 6% of conversations; on ABCD it reports .657 and is overconfident there (ECE .19). Workflows unseen in training are much weaker. A Turkish sibling, Lavoir-TR, is built on MoganBERT-TR. The weights are CC BY-NC 4.0 because some training data is non-commercial, so they are not mirrored or packaged here; the code is Apache-2.0.

## What it decides

- **choice** — picks one option from a set
- **score** — places the input on an ordered scale
- **noul** — answers a yes/no question with one probability
- **route** — sends the input to one of several destinations

## At a glance

| | |
|---|---|
| Base model | `answerdotai/ModernBERT-large` |
| Maker | Mogan AI |
| Released | 2026-09-25 |
| License | cc-by-nc-4.0 |
| Reported accuracy | 94.2% |
| Reported latency | 31 ms median, 142 ms p95 per question on one NVIDIA GH200 (bf16) |

## Get the weights

```bash
pip install systemonemodels
systemone pull moganai/lavoir
```

The files are served from the maker's Hugging Face repository, [`moganai/lavoir`](https://huggingface.co/moganai/lavoir), and verified against the checksums recorded here.

## Read more

- [Model card](https://huggingface.co/moganai/lavoir)
- [Code](https://github.com/moganai/lavoir)
- [Paper (arXiv 2609.30706)](https://arxiv.org/abs/2609.30706)
- [Lavoir-TR (Turkish)](https://huggingface.co/moganai/lavoir-tr)

---

*This page was opened by System One for Mogan AI, who can claim the organisation and take it over at any time.*

---

From System One Models — https://systemonemodels.ai/ · every System One model: https://systemonemodels.ai/system-one-models