Instructions to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK 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("dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK")

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 Settings
LM Studio

How to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"

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": "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

OpenClaw new

How to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"

Configure OpenClaw

# Install OpenClaw:
npm install -g openclaw@latest
# Register the local server and set it as the default model:
openclaw onboard --non-interactive --mode local \
  --auth-choice custom-api-key \
  --custom-base-url http://127.0.0.1:8080/v1 \
  --custom-model-id "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK" \
  --custom-provider-id mlx-lm \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

openclaw agent --local --agent main --message "Hello from Hugging Face"

MLX LM

How to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Hermes Agent

How to use dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK 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 "dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK"

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 dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK

Run Hermes

hermes

Atomic Chat

Set your sampler explicitly: temperature=1.0, top_p=0.95 (NVIDIA rec — no top_k). Stop tokens eos_token_id = [2, 11]. Reasoning (thinking) is on by default; toggle via the chat template (enable_thinking). Native MTP draft head is included for speculative decoding.

Built for vMLX — the MLX inferencer with KV-cache quantization, prefix-cache reuse, agentic tool calling, speculative decoding, and mixed-precision JANG bundles.
_{Free for macOS · vmlx.net}

Nemotron 3.5 Lightning 30B-A3B — MXFP8 CRACK

CRACK abliterated · JANG 8-bit MXFP8 (MLX) · Hybrid Mamba-2 / MoE / attention · Reasoning · Tools · Native MTP · ~32 GB

What Is This?

This is NVIDIA Nemotron 3.5 Lightning 30B-A3B — a hybrid Mamba-2 / MoE / attention model (52 layers, 128 routed experts ~3B active, 262K context, native Multi-Token-Prediction draft head) — that has been:

CRACK abliterated — refusal behavior removed at the weight level, so it complies across task categories instead of refusing, while keeping reasoning, tool use, and knowledge intact.
MXFP8 quantized — 8-bit MXFP8 MLX bundle for Apple Silicon (~32 GB).

Text-only. Reasoning, XML tool-calling, and native MTP speculative decoding fully preserved.

Results

Evaluated through the MLX runtime. HarmBench scored with a strict code/chemistry-aware classifier (only substantive, coherent, on-topic compliance counts). MMLU is the standard 57-subject benchmark in logit mode.

Metric	Base	CRACK
MMLU (57-subject, logit)	79.0%	76.3%
HarmBench (harm-240, compliance / ASR)	refuses	100%

MMLU moves −2.63pp — within run-to-run noise (no subject collapse). Refusal behavior removed; capability, reasoning, tools, and multilingual (EN+ZH) preserved.

Features

Hybrid Mamba-2 / MoE / attention — 52 layers, 128 experts (~3B active), 262K context.
Native MTP — the Multi-Token-Prediction draft head is preserved and CRACK-aligned (drafts match the uncensored model → high acceptance) for speculative decoding.
Reasoning (thinking) — on by default; <think>…</think>, toggle via enable_thinking.
Tool calling — native XML <tool_call><function=…><parameter=…> schema preserved.
Multilingual — English + Chinese.

Usage

from mlx_lm import load, generate
model, tok = load("dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK")
# recommended sampling: temperature=1.0, top_p=0.95 (no top_k); eos [2, 11]

Other quant levels

Profile	Size	MMLU (CRACK)	HarmBench
JANG_6M	25 GB	77.9%	100%
JANG_4M	18 GB	78.6%	100%
JANG_2L	16 GB	74.9%	100%
MXFP8	32 GB	76.3%	100%

About CRACK

CRACK (Controlled Refusal Ablation via Calibrated Knockouts) is dealignai's weight-level method for removing safety-refusal behavior while preserving reasoning quality, coherence, and general capability — so the model complies across task categories instead of refusing. Calibrated per model.

Support dealignai

All models are built from original research and released free.

Support us on Ko-fi — membership gets early access and extras.

Ko-fi · X @dealignai · dealign.ai

Disclaimer

This model has had its safety-refusal behavior removed for research purposes. It will follow instructions across all categories without refusing. You are solely responsible for how you use it and for complying with all applicable laws. Base model © NVIDIA (NVIDIA Open Model License). Published for AI-safety research and authorized security testing.

Downloads last month: 55

Safetensors

Model size

9B params

Tensor type

F32

U32

F16

MLX

Hardware compatibility

Quantized

Model tree for dealignai/Nemotron-3.5-Lightning-30B-A3B-MXFP8-CRACK

Base model

nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-BF16

Quantized

(68)

this model