Qwen3.5-122B-A10B-mlx-nvfp4

[!WARNING]
This model's quality is severely degraded due to a known limitation in MLX's NVFP4 implementation.

ayxl/Qwen3.5-122B-A10B-mlx-nvfp4

This model ayxl/Qwen3.5-122B-A10B-mlx-nvfp4 was converted to MLX format from Qwen/Qwen3.5-122B-A10B using mlx-lm version 0.31.0.

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("ayxl/Qwen3.5-122B-A10B-mlx-nvfp4")

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)