mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-mxfp4

This model mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-mxfp4 was converted to MLX format from nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16 using mlx-lm version 0.31.3 (MXFP4, group size 32).

NVIDIA-Nemotron-3.5-Lightning-30B-A3B is a hybrid Mamba-2 / attention Mixture-of-Experts model (~31B total parameters, ~3B active per token).

Use with mlx

pip install mlx-lm

from mlx_lm import load, generate

model, tokenizer = load("mlx-community/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-mxfp4")

prompt = "Hello, who are you?"

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)