Qwen3-Next-80B-A3B-Instruct-mxfp4-mlx

The prompt to test this:

mlx_lm.generate --model Qwen3-Next-80B-A3B-Instruct-mxfp4-mlx --max-tokens 20000 \
--prompt "Describe in less than 100 words what would happen if a black hole would appear around Uranus"

==========
If a black hole appeared near Uranus, its immense gravity would instantly disrupt the planet’s orbit,
tearing it apart through tidal forces. Surrounding moons, rings, and nearby objects would be swallowed
or flung into space. The accretion disk would emit intense X-rays, vaporizing nearby ice and gas.
Earth would detect the sudden gamma-ray burst and gravitational waves, but the solar system’s stability
would be shattered—Uranus would vanish, and the Sun’s gravitational balance altered,
potentially destabilizing other planets. Life on Earth would face catastrophic radiation and orbital chaos.
==========
Prompt: 29 tokens, 149.991 tokens-per-sec
Generation: 117 tokens, 57.075 tokens-per-sec
Peak memory: 42.672 GB

"Describe in less than 10 words what would happen if a black hole would appear around Uranus" 
==========
Uranus swallowed; solar system disrupted instantly.
==========

This model Qwen3-Next-80B-A3B-Instruct-mxfp4-mlx was converted to MLX format from Qwen/Qwen3-Next-80B-A3B-Instruct using mlx-lm version 0.27.1.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("Qwen3-Next-80B-A3B-Instruct-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)