Qwen3-32B-Element5-Heretic-qx86-hi-mlx

Brainwave: 0.483,0.596,0.738,0.754,0.394,0.802,0.710

This is a nuslerp merge of the following models:

• Skywork/MindLink-32B-0801 (Engineer4)
• Akicou/DeepKAT-32B (Engineer4)
• microsoft/FrogBoss-32B-2510 (Element3)
• ValiantLabs/Qwen3-32B-Guardpoint (Element4)
• ReadyArt/Dark-Nexus-32B-v2.0 (Element5)

The ReadyArt/Dark-Nexus-32B-v2.0 was included for its rich vocabulary--more about that on the model card :)

The model has been abliterated with Heretic by DavidAU

Refusals: 94/100
Final: 27/100 // 0.1708 KLD

Metrics are pending

-G

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("Qwen3-32B-Element5-Heretic-qx86-hi-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, return_dict=False,
    )

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