Qwen3-8B-NotSure-mxfp4-mlx

8B or not 8B—that is the question

NotSure, the ultimate generalist, a mind-meld of great models in the field

• Not for coding
• Not for reasoning
• Its scores are abysmal
• It thinks a lot and changes its mind
• Let loose on a task it thinks of itself as a mind meld of Sisko and Nog

In uncertain times, the best decisions are left up to chance.

-G

--for a more confidently not sure version, try the qx86-hi

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("Qwen3-8B-NotSure-mxfp4-mlx")

prompt = "hello"

if tokenizer.chat_template is not None:
    messages = [{"role": "user", "content": prompt}]
    prompt = tokenizer.apply_chat_template(
        messages, add_generation_prompt=True
    )

response = generate(model, tokenizer, prompt=prompt, verbose=True)