# Amir Mahdi Madani: decima

> A 122M-parameter late-interaction decision model on multilingual-e5-small: choose, verify, ordered score and rank with a probability per option, int8 ONNX at about 20 ms on one CPU core, the same answer in any option order, and a local server in the System One wire format.

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

## Facts

| | |
|---|---|
| Maker | Amir Mahdi Madani (https://systemonemodels.ai/amyrmahdy) |
| Decides | choice, score, noul, rank, classify, route |
| Architecture | decima |
| Base model | intfloat/multilingual-e5-small |
| Parameters | 122M |
| Licence | apache-2.0 |
| Availability | Open weights |
| Released | 2026-09-27 |
| Latest version | 1.1.0 |

## Reported evaluation

Suite: jabr/classifier-benchmark v2 (49 tasks, 866 cases, macro accuracy), the maker's own run with int8 ONNX on CPU. Numbers are the publisher's own.

- Decision accuracy: 61.6%

- Median latency: 20 ms
- p95 latency: 23 ms

## Model card

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

# Decima-small 1.1

A 122M-parameter late-interaction decision model on multilingual-e5-small: choose, verify, ordered score and rank with a probability per option, int8 ONNX at about 20 ms on one CPU core, the same answer in any option order, and a local server in the System One wire format.

A personal release by Amir Mahdi Madani. The state and question are encoded once, each free-text option is encoded separately and cross-attends to the state, and the model gives one logit per option, so option order cannot change the answer. One shipped temperature, fitted on a hold-out of the training mix. On the community jabr/classifier-benchmark v2 (a third-party suite) the maker's own int8 run gives 0.616 macro accuracy; score is the weakest type. Ships int8 and fp32 ONNX exports and PyTorch weights; runs in the browser. Siblings decima-base and decima-agent.

## 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
- **rank** — orders a set of items
- **classify** — assigns a category from a fixed taxonomy
- **route** — sends the input to one of several destinations

## At a glance

| | |
|---|---|
| Parameters | 122M |
| Base model | `intfloat/multilingual-e5-small` |
| Maker | Amir Mahdi Madani |
| Released | 2026-09-27 |
| License | apache-2.0 |
| Reported accuracy | 61.6% |
| Reported latency | 20 ms p50 / 23 ms p95 on one Intel Core Ultra 7 P-core, ONNX Runtime int8, 4 options |

## Get the weights

```bash
pip install systemonemodels
systemone pull amyrmahdy/decima
```

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

## Read more

- [Model card](https://huggingface.co/amyrmahdy/decima-small)
- [decima-base](https://huggingface.co/amyrmahdy/decima-base)
- [Code](https://github.com/amyrmahdy/decima)

---

*This page was opened by System One for Amir Mahdi Madani, 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