# Akiv (akivet): mica

> A Qwen3.5-4B decision model with a merged LoRA, trained on English and Korean. It reads the state once and returns a probability per answer (yes/no, 2 to 255 options, or a 2 to 10 level score) from single-token label logits, with no generated text.

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

## Facts

| | |
|---|---|
| Maker | Akiv (akivet) (https://systemonemodels.ai/akivet) |
| Decides | choice, score, noul |
| Architecture | mica |
| Base model | qwen/qwen3.5-4b |
| Parameters | 4.0B |
| Context | 8K tokens |
| Licence | apache-2.0 |
| Availability | Open weights |
| Released | 2026-09-25 |
| Latest version | 0.1.0 |

## Reported evaluation

Suite: Maker's held-out set (7,328 English and Korean items written after the training data was frozen). Numbers are the publisher's own.

- Decision accuracy: 67.4%
- Calibration error (ECE): 0.054
- Median latency: 54 ms
- p95 latency: 552 ms

## Model card

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

# Mica v0.1 4B

A Qwen3.5-4B decision model with a merged LoRA, trained on English and Korean. It reads the state once and returns a probability per answer (yes/no, 2 to 255 options, or a 2 to 10 level score) from single-token label logits, with no generated text.

A personal release by Akiv. A merged rank-16 LoRA on all 32 layers with no added heads; several questions about one state share the prefill. Calibration is a single temperature (1.124) fitted on the maker's calibration set. Serves a Jev-compatible /v1/systemone endpoint through llama.cpp, with the maker's own GGUF builds in the same repo. The maker's run on JevBench's public items is not on that third-party board, and part of the hard tier was used during development. Weak on knowledge-heavy questions (MMLU-Pro 53.0 in the maker's run); instructions planted inside the state still sway answers; inputs over 8,192 tokens are rejected.

## 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

## At a glance

| | |
|---|---|
| Parameters | 4B |
| Base model | `Qwen/Qwen3.5-4B` |
| Maker | Akiv (akivet) |
| Released | 2026-09-25 |
| License | apache-2.0 |
| Reported accuracy | 67.4% |
| Reported latency | 54 ms p50 / 552 ms p95 on one RTX 3090, BF16 GGUF via llama.cpp, on JevBench's public items |

## Get the weights

```bash
pip install systemonemodels
systemone pull akivet/mica
```

The files are served from the maker's Hugging Face repository, [`sky7350/Mica-v0.1-4B`](https://huggingface.co/sky7350/Mica-v0.1-4B), and verified against the checksums recorded here.

## Read more

- [Model card](https://huggingface.co/sky7350/Mica-v0.1-4B)
- [Code and server](https://github.com/akivet/Mica-v0.1-4B)

---

*This page was opened by System One for Akiv (akivet), 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