# Metask Lab: metask-jev

> Metask Lab's calibrated decision model in 16 languages. Qwen3.5-4B with a merged LoRA trained with the Nimble candidate-logit objective; one forward pass and a softmax over at most 26 answer-letter logits, with one temperature per question type.

- Page: https://systemonemodels.tech/metask-lab/metask-jev
- API: https://api.systemonemodels.tech/v1/models/metask-lab/metask-jev
- Download: `pip install systemonemodels && systemone pull metask-lab/metask-jev`

## Facts

| | |
|---|---|
| Maker | Metask Lab (https://systemonemodels.tech/metask-lab) |
| Decides | choice, score, noul, classify, route |
| Architecture | metask-jev |
| Base model | qwen/qwen3.5-4b |
| Parameters | 4.5B |
| Context | 4K tokens |
| Licence | apache-2.0 |
| Availability | Open weights |
| Released | 2026-09-21 |
| Latest version | 2026.09 |

## Reported evaluation

Suite: JevBench v1.2 public set (231 items) at 4,096 tokens, maker's run. Numbers are the publisher's own.

- Decision accuracy: 80.1%

- Median latency: 62.8 ms

## Model card

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

# Metask-Jev-4B

Metask Lab's calibrated decision model in 16 languages. Qwen3.5-4B with a merged LoRA trained with the Nimble candidate-logit objective; one forward pass and a softmax over at most 26 answer-letter logits, with one temperature per question type.

Trained on 60.9k decisions: human-labelled public data, 16.1k MASSIVE routing items in 14 locales, and 390 synthetic policy items that deliberately mirror JevBench's hard-tier families (disclosed). The maker reports 80.1% on JevBench v1.2's 231 public items at 4,096 tokens, a 78.9% macro on 13 human-labelled subsets, 85.9% on held-out MASSIVE in 14 locales, and ECE 0.114 falling to 0.040 after temperature fitting. Its claim that it would rank first is its own estimate: the official JevBench v1.4.2 run placed it 12th (79.7% public, 27.6% sealed). Validated at 4,096 tokens; the base supports 262,144.

## 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 calibrated probability
- **classify** — assigns a category from a fixed taxonomy
- **route** — sends the input to one of several destinations

## At a glance

| | |
|---|---|
| Parameters | 4.5B |
| Base model | `qwen/qwen3.5-4b` |
| Maker | Metask Lab |
| Released | 2026-09-21 |
| License | apache-2.0 |
| Reported accuracy | 80.1% |
| Reported latency | 62.8 ms p50 on an RTX 4090 |

## Get the weights

```bash
pip install systemonemodels
systemone pull metask-lab/metask-jev
```

The files are served from the maker's Hugging Face repository, [`wayfind/metask-jev-4b-policy-mix`](https://huggingface.co/wayfind/metask-jev-4b-policy-mix), and verified against the checksums recorded here.

## Read more

- [Model card](https://huggingface.co/wayfind/metask-jev-4b-policy-mix)
- [Code and recipe](https://github.com/metask-ai/metask-jev)

---

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

---

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