Special-purpose model. This checkpoint does exactly one job: cleaning up speech-to-text output under fixed system prompts and a strict JSON output contract ({"result": "..."}). It is not a general-purpose chat or instruction model; off-task use (open-ended chat, question answering) is unsupported and will underperform the base model. All Yooz task models are collected in Yooz Working Models; for general-purpose local LLMs see Lean QAT MLX.

Pre-release v2.1 iteration of the Yooz Quality speech-to-text touchup tier from Yooz Labs: a fused LoRA fine-tune of Qwen3.5-0.8B (4-bit MLX). The shipped, fully documented tier is Yooz-Quality-v2; see that card for the task description, the exact system prompts (use them verbatim), the output contract, and evaluation methodology, all of which apply unchanged to v2.1. Evaluation of the v2.1 line is in progress and this checkpoint may be updated.

Converted to MLX format from mlx-community/Qwen3.5-0.8B-MLX-4bit using mlx-lm version 0.31.3.

Use with mlx

python

import json
from mlx_lm import load, generate

model, tokenizer = load("YoozLabs/Yooz-Quality-v2.1-Qwen3.5-0.8B")

PROOFREAD_SYSTEM = (
    "You are a copy editor. Proofread the speech-to-text transcription provided by the "
    "user. Fix spelling, punctuation, capitalization, and obvious word-choice errors "
    "caused by the speech-to-text engine. Convert spoken numbers to digits where it "
    "improves clarity (e.g. \"nine am\" -> \"9 AM\"). Keep the speaker's voice; do NOT "
    "rephrase or remove filler words (um, uh, like). "
    "Output ONLY a single JSON object of the form {\"result\": \"<corrected>\"} where "
    "<corrected> is the actual corrected text. No explanation, no markdown."
)

raw_stt = "um so the meeting is at nine am tomorow with sarah from product"

prompt = tokenizer.apply_chat_template(
    [{"role": "system", "content": PROOFREAD_SYSTEM},
     {"role": "user", "content": raw_stt}],
    add_generation_prompt=True, tokenize=False,
)
response = generate(model, tokenizer, prompt=prompt, max_tokens=320)
print(json.loads(response)["result"])
# Um, so the meeting is at 9 AM tomorrow with Sarah from Product.

License

Apache 2.0, matching the Qwen3.5 base model.

Contact

Questions: dev@yooz.info
Bug reports: github.com/yooz-labs/yooz-engine