d1-2B-OptiQ-4bit

LiquidAI/d1-2B, a decision model for Apple Silicon: it answers typed questions at POST /v1/systemone in optiq serve.

decision

On disk1.7 GB

Memoryabout 3 GB

Use it

Install

pip install mlx-optiq

Serve it

optiq serve turns on POST /v1/systemone by itself when the model carries a decision head. Decision models docs.

optiq serve --model mlx-community/d1-2B-OptiQ-4bit

Ask it a question

curl http://127.0.0.1:8080/v1/systemone -d '{
  "state": "Checkout has been failing for every customer for the last hour.",
  "questions": {
    "urgent": {"type": "noul", "instructions": "Is this urgent?"},
    "team": {"type": "choice", "criteria": {"billing": "Payments", "technical": "Outages and errors"}}
  }
}'

From Python

from optiq.decision import DecisionModel

model = DecisionModel.load("mlx-community/d1-2B-OptiQ-4bit")

Details

Family	Decision models · Decision models docs
Architecture	LFM3-VL
Base model	LiquidAI/d1-2B
Size on disk	1.7 GB
Memory needed	about 3 GB
Downloads	0 in the last month
Updated	2026-10-30
Hugging Face	mlx-community/d1-2B-OptiQ-4bit