Instructions to use RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 with MLX:

# Make sure mlx-lm is installed
# pip install --upgrade mlx-lm

# Generate text with mlx-lm
from mlx_lm import load, generate

model, tokenizer = load("RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4")

prompt = "Write a story about Einstein"
messages = [{"role": "user", "content": prompt}]
prompt = tokenizer.apply_chat_template(
    messages, add_generation_prompt=True
)

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

Notebooks
Google Colab
Kaggle
Local Apps
LM Studio

Pi new

How to use RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4"

Configure the model in Pi

# Install Pi:
npm install -g @mariozechner/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "mlx-lm": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 with Hermes Agent:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4"

Configure Hermes

# Install Hermes:
curl -fsSL https://hermes-agent.nousresearch.com/install.sh | bash
hermes setup
# Point Hermes at the local server:
hermes config set model.provider custom
hermes config set model.base_url http://127.0.0.1:8080/v1
hermes config set model.default RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4

Run Hermes

hermes

MLX LM

How to use RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Edit: 2026-03-25

Recommend Quantized Model: RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-oQ3.5

This model was converted to MLX format from nvidia/Nemotron-Cascade-2-30B-A3B using mlx-lm version 0.30.7.

Conversion Command

$ uv run mlx_lm.convert --hf-path nvidia/Nemotron-Cascade-2-30B-A3B --mlx-path Nemotron-Cascade-2-30B-A3B-mlx-mxfp4 -q --q-mode mxfp4 -q-group-size 32

Benchmark

Model	File size	MMLU	JMMLU	HELLASWAG	ARC_CHALLENGE	GSM8K
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	15.65 GB	45.7%	33.7%	28.3%	72.3%	82.3%
Nemotron-Cascade-2-30B-A3B-oQ3.5	13.32 GB	65.7%	61.0%	76.0%	85.7%	89.7%
Nemotron-Cascade-2-30B-A3B-oQ4	16.91 GB	68.3%	61.7%	-	-	-

Detail

Model	Benchmark	Accuracy	Correct	Total	Time(s)
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	MMLU	45.7%	137	300	946.7
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	JMMLU	33.7%	101	300	558.5
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	HELLASWAG	28.3%	85	300	647
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	GSM8K	82.3%	247	300	1781.6
Nemotron-Cascade-2-30B-A3B-mlx-mxfp4	ARC_CHALLENGE	72.3%	217	300	462.5
Nemotron-Cascade-2-30B-A3B-oQ3.5	MMLU	65.7%	197	300	696
Nemotron-Cascade-2-30B-A3B-oQ3.5	JMMLU	61.0%	183	300	294.7
Nemotron-Cascade-2-30B-A3B-oQ3.5	HELLASWAG	76.0%	228	300	314.4
Nemotron-Cascade-2-30B-A3B-oQ3.5	GSM8K	89.7%	269	300	992.6
Nemotron-Cascade-2-30B-A3B-oQ3.5	ARC_CHALLENGE	85.7%	257	300	204.5
Nemotron-Cascade-2-30B-A3B-oQ4	MMLU	68.3%	205	300	572.7
Nemotron-Cascade-2-30B-A3B-oQ4	JMMLU	61.7%	185	300	239.2

Downloads last month: 41

Safetensors

Model size

32B params

Tensor type

U32

BF16

F32

MLX

Hardware compatibility

4-bit

Model tree for RepublicOfKorokke/Nemotron-Cascade-2-30B-A3B-mlx-mxfp4

Base model

nvidia/Nemotron-Cascade-2-30B-A3B

Quantized

(30)

this model