Instructions to use majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit 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("majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit")

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 majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit with Pi:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"

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": "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Hermes Agent new

How to use majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit 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 "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"

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 majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit

Run Hermes

hermes

OpenClaw new

How to use majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit with OpenClaw:

Start the MLX server

# Install MLX LM:
uv tool install mlx-lm
# Start a local OpenAI-compatible server:
mlx_lm.server --model "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"

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 "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit" \
  --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 majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit with MLX LM:

Generate or start a chat session

# Install MLX LM
uv tool install mlx-lm
# Interactive chat REPL
mlx_lm.chat --model "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"

Run an OpenAI-compatible server

# Install MLX LM
uv tool install mlx-lm
# Start the server
mlx_lm.server --model "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit"
# Calling the OpenAI-compatible server with curl
curl -X POST "http://localhost:8000/v1/chat/completions" \
   -H "Content-Type: application/json" \
   --data '{
     "model": "majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit",
     "messages": [
       {"role": "user", "content": "Hello"}
     ]
   }'

Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit / README.md

majentik

docs: upstream-first KV-cache guidance (q8_0/q4_0, mainline Hadamard rotation); fork demoted to experimental

f7c2f81 verified 4 days ago

preview code

Raw

History Blame Contribute Delete

14.1 kB

	---
	license: other
	license_name: nvidia-open-model-license
	license_link: https://developer.download.nvidia.com/licenses/nvidia-open-model-license-agreement-june-2024.pdf
	base_model: nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16
	tags: [nemotron, multimodal, mamba2, moe, quantized, turboquant, mlx, apple-silicon,
	mlx-lm, text-tower-only]
	library_name: mlx
	pipeline_tag: text-generation
	language: [en]
	datasets: [nvidia/Nemotron-Image-Training-v3]
	inference: false
	---

	> [!TIP]
	> KV-cache quantization without any fork (recommended, 2026): upstream
	> llama.cpp/Ollama now cover this natively — use `-ctk q8_0 -ctv q8_0`
	> (~half KV memory, negligible quality loss: perplexity +0.002–0.05) or
	> `-ctk q4_0 -ctv q4_0` (~quarter memory, ≈7.6% perplexity increase). In
	> Ollama: `OLLAMA_KV_CACHE_TYPE=q8_0` with `OLLAMA_FLASH_ATTENTION=1`. Keep
	> K and V types symmetric to stay on the fast fused Flash-Attention path.
	> Since April 2026, mainline llama.cpp also applies Hadamard rotation to
	> KV activations ([PR #21038](https://github.com/ggml-org/llama.cpp/pull/21038)),
	> which greatly improves low-bit KV quality (opt-out:
	> `LLAMA_ATTN_ROT_DISABLE=1`).
	>
	> The RotorQuant/TurboQuant fork flow below is experimental/legacy: the
	> TurboQuant llama.cpp PR was closed without merging (June 2026) and the fork
	> is unmaintained relative to mainline. It is NOT required to use this model.
	<!-- kv-upstream-note -->

	# Nemotron-3-Nano-Omni-30B-A3B-Reasoning - TurboQuant MLX 3-bit

	MLX 3-bit quantization of the text tower of `Nemotron-3-Nano-Omni-30B-A3B-Reasoning` (`nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16`)
	with TurboQuant weight method. Apple Silicon native via `mlx-lm`.

	This variant covers the LLM backbone only. Vision (CRADIO v4-H) + audio (Parakeet-TDT-0.6B-v2)
	encoders are NOT included — MLX-VLM Nemotron-Omni model class is pending upstream support
	(no PR observed as of 2026-05-04). For multimodal inference, use the GGUF variants with
	`llama-mtmd-cli` instead.

	For the matched-KV stack — TurboQuant weights + TurboQuant KV-cache modifier —
	see [`majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit-TQ-KV`](https://huggingface.co/majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit-TQ-KV).
	For the runtime KV-cache modifier itself, see
	[`majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant`](https://huggingface.co/majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant).

	## Quickstart

	```python
	# Today (mlx-lm 0.31.x): the NemotronH_Nano_Omni_Reasoning_V3 model class
	# is not yet registered in mlx-lm. The cell below is the API shape that WILL
	# work once upstream lands the class (track ml-explore/mlx-lm#386).

	from mlx_lm import load, generate

	model, tokenizer = load("majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit")

	prompt = tokenizer.apply_chat_template(
	[{"role": "user", "content": "Solve: 17 * 23"}],
	add_generation_prompt=True,
	enable_thinking=False, # set True to enable extended reasoning (default)
	)

	response = generate(
	model, tokenizer,
	prompt=prompt,
	max_tokens=512,
	sampler=lambda x: x.argmax(axis=-1), # or use mlx_lm.sample_utils.make_sampler(temp=0.6, top_p=0.95)
	)
	print(response)
	```

	> ⚠️ This variant covers the text tower only. For multimodal inference (vision + audio + video), use the GGUF variants with `llama-mtmd-cli` — see the GGUF cards in this family.

	## Modality matrix

	\| Modality \| Encoder \| Quantization in this variant \|
	\|---\|---\|---\|
	\| Text \| LLM backbone (Mamba-2 + Transformer hybrid Sparse MoE) \| per the variant suffix \|
	\| Image \| CRADIO v4-H \| BF16 (kept full-precision in every non-GGUF variant; GGUF uses mmproj-F16 split file) \|
	\| Audio \| Parakeet-TDT-0.6B-v2 \| BF16 (same rationale) \|
	\| Video \| Parakeet-TDT-0.6B-v2 + frame sampler \| BF16 (≤ 2 min, 256 frames @ 2 FPS) \|

	NVIDIA's official FP8 / NVFP4 recipe keeps both encoders + the cross-modal
	MLP projectors in BF16 to preserve multimodal accuracy. We follow that
	convention in every quantized variant we ship.

	## Runtime quirks

	### MLX-LM (text-only)

	This variant covers the LLM backbone only. Vision + audio encoders
	are NOT included — MLX-VLM Nemotron-Omni model class is
	pending upstream support (no PR observed as of 2026-05-04).

	Use the `mlx_lm.generate` API; `enable_thinking` is a runtime flag
	(see below).

	### Reasoning mode

	`enable_thinking` defaults to `True`. To disable extended reasoning
	(e.g., for latency-sensitive cases), pass `enable_thinking=False`
	to the chat template / generate call. No separate "no-think"
	variant card exists — this is a runtime flag, not a model variant.

	## Quant trade-off (MLX lane)

	\| Bits \| Approx size \| Use case \| Recommendation \|
	\|---\|---\|---\|---\|
	\| 2-bit \| ~8.1 GB \| Aggressive quantization \| Very low-RAM Macs \|
	\| 3-bit \| ~11 GB \| Lossy but small \| Low-RAM Macs \|
	\| 4-bit \| ~13 GB \| Balanced default \| Recommended for most Macs \|
	\| 5-bit \| ~16 GB \| Higher fidelity \| Quality-sensitive \|
	\| 6-bit \| ~19 GB \| Approaching FP16 quality \| High-fidelity \|
	\| 8-bit \| ~24 GB \| Near-lossless reference \| Fidelity-critical work \|

	(Current variant — 3bit — is bolded.)

	## Variants in this family

	(Showing 56 sibling variants under `majentik/nemotron3-nano-omni-30b-`. The current variant — `TurboQuant-MLX-3bit` — is bolded*.)

	\| Variant \| Runtime \| Approx size \| Use case \|
	\|---\|---\|---\|---\|
	\| [mmproj-F16](https://huggingface.co/majentik/nemotron3-nano-omni-30b-mmproj-f16) \| llama-mtmd-cli \| ~1-2 GB \| Multimodal projector (pair with any GGUF) \|
	\| [RotorQuant](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant) \| runtime modifier \| n/a \| KV-cache root (weight-agnostic) \|
	\| [RotorQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-IQ4_XS) \| llama.cpp \| ~26 GB \| Lossy 4-bit, low-RAM CPU/edge \|
	\| [RotorQuant-GGUF-MXFP4_MOE](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-MXFP4_MOE) \| llama.cpp \| ~30 GB \| MXFP4 MoE quant \|
	\| [RotorQuant-GGUF-Q2_K](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q2_K) \| llama.cpp \| ~18 GB \| Lossy, low-RAM CPU/edge \|
	\| [RotorQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q3_K_M) \| llama.cpp \| ~23 GB \| Smaller 3-bit, CPU-friendly \|
	\| [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q4_K_M) \| llama.cpp \| ~33 GB \| Balanced default \|
	\| [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q5_K_M) \| llama.cpp \| ~40 GB \| Higher fidelity, more RAM \|
	\| [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q8_0) \| llama.cpp \| ~63 GB \| Near-lossless reference \|
	\| [RotorQuant-GGUF-IQ4_XS-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-iq4_xs-rq-kv) \| llama.cpp \| ~26 GB \| IQ4_XS + RotorQuant KV \|
	\| [RotorQuant-GGUF-MXFP4_MOE-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-mxfp4_moe-rq-kv) \| llama.cpp \| ~30 GB \| MXFP4 MoE + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q2_K-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q2_k-rq-kv) \| llama.cpp \| ~18 GB \| Q2_K + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q3_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q3_k_m-rq-kv) \| llama.cpp \| ~23 GB \| Q3_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q4_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q4_k_m-rq-kv) \| llama.cpp \| ~33 GB \| Q4_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q5_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q5_k_m-rq-kv) \| llama.cpp \| ~40 GB \| Q5_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q8_0-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q8_0-rq-kv) \| llama.cpp \| ~63 GB \| Q8_0 + RotorQuant KV \|
	\| [RotorQuant-MLX-2bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-2bit) \| mlx-lm \| ~9.6 GB \| Apple Silicon, smallest \|
	\| [RotorQuant-MLX-2bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-2bit-rq-kv) \| mlx-lm \| ~9.6 GB \| 2-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-3bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-3bit) \| mlx-lm \| ~14 GB \| Apple Silicon, small \|
	\| [RotorQuant-MLX-3bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-3bit-rq-kv) \| mlx-lm \| ~14 GB \| 3-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-4bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-4bit) \| mlx-lm \| ~19 GB \| Apple Silicon balanced \|
	\| [RotorQuant-MLX-4bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-4bit-rq-kv) \| mlx-lm \| ~19 GB \| 4-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-5bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-5bit) \| mlx-lm \| ~23 GB \| Apple Silicon, higher fidelity \|
	\| [RotorQuant-MLX-5bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-5bit-rq-kv) \| mlx-lm \| ~23 GB \| 5-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-6bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-6bit) \| mlx-lm \| ~27 GB \| Apple Silicon, near-lossless \|
	\| [RotorQuant-MLX-6bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-6bit-rq-kv) \| mlx-lm \| ~27 GB \| 6-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-8bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-8bit) \| mlx-lm \| ~35 GB \| Apple Silicon reference \|
	\| [RotorQuant-MLX-8bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-8bit-rq-kv) \| mlx-lm \| ~35 GB \| 8-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-MXFP4](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-mxfp4) \| mlx-lm \| ~19 GB \| Apple Silicon MXFP4 \|
	\| [TurboQuant](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant) \| runtime modifier \| n/a \| KV-cache root (weight-agnostic) \|
	\| [TurboQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-IQ4_XS) \| llama.cpp \| ~26 GB \| Lossy 4-bit, low-RAM CPU/edge \|
	\| [TurboQuant-GGUF-MXFP4_MOE](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-MXFP4_MOE) \| llama.cpp \| ~30 GB \| MXFP4 MoE quant \|
	\| [TurboQuant-GGUF-Q2_K](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q2_K) \| llama.cpp \| ~18 GB \| Lossy, low-RAM CPU/edge \|
	\| [TurboQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q3_K_M) \| llama.cpp \| ~23 GB \| Smaller 3-bit, CPU-friendly \|
	\| [TurboQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q4_K_M) \| llama.cpp \| ~33 GB \| Balanced default \|
	\| [TurboQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q5_K_M) \| llama.cpp \| ~40 GB \| Higher fidelity, more RAM \|
	\| [TurboQuant-GGUF-Q8_0](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q8_0) \| llama.cpp \| ~63 GB \| Near-lossless reference \|
	\| [TurboQuant-GGUF-IQ4_XS-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-iq4_xs-tq-kv) \| llama.cpp \| ~26 GB \| IQ4_XS + TurboQuant KV \|
	\| [TurboQuant-GGUF-MXFP4_MOE-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-mxfp4_moe-tq-kv) \| llama.cpp \| ~30 GB \| MXFP4 MoE + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q2_K-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q2_k-tq-kv) \| llama.cpp \| ~18 GB \| Q2_K + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q3_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q3_k_m-tq-kv) \| llama.cpp \| ~23 GB \| Q3_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q4_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q4_k_m-tq-kv) \| llama.cpp \| ~33 GB \| Q4_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q5_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q5_k_m-tq-kv) \| llama.cpp \| ~40 GB \| Q5_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q8_0-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q8_0-tq-kv) \| llama.cpp \| ~63 GB \| Q8_0 + TurboQuant KV \|
	\| [TurboQuant-MLX-2bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-2bit) \| mlx-lm \| ~9.6 GB \| Apple Silicon, smallest \|
	\| [TurboQuant-MLX-2bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-2bit-tq-kv) \| mlx-lm \| ~9.6 GB \| 2-bit + TurboQuant KV \|
	\| TurboQuant-MLX-3bit \| mlx-lm \| ~14 GB \| Apple Silicon, small \|
	\| [TurboQuant-MLX-3bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-3bit-tq-kv) \| mlx-lm \| ~14 GB \| 3-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-4bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-4bit) \| mlx-lm \| ~19 GB \| Apple Silicon balanced \|
	\| [TurboQuant-MLX-4bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-4bit-tq-kv) \| mlx-lm \| ~19 GB \| 4-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-5bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-5bit) \| mlx-lm \| ~23 GB \| Apple Silicon, higher fidelity \|
	\| [TurboQuant-MLX-5bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-5bit-tq-kv) \| mlx-lm \| ~23 GB \| 5-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-6bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-6bit) \| mlx-lm \| ~27 GB \| Apple Silicon, near-lossless \|
	\| [TurboQuant-MLX-6bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-6bit-tq-kv) \| mlx-lm \| ~27 GB \| 6-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-8bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-8bit) \| mlx-lm \| ~35 GB \| Apple Silicon reference \|
	\| [TurboQuant-MLX-8bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-8bit-tq-kv) \| mlx-lm \| ~35 GB \| 8-bit + TurboQuant KV \|

	---
	license: other
	license_name: nvidia-open-model-license
	license_link: https://developer.download.nvidia.com/licenses/nvidia-open-model-license-agreement-june-2024.pdf
	base_model: nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16
	tags: [nemotron, multimodal, mamba2, moe, quantized, turboquant, mlx, apple-silicon,
	mlx-lm, text-tower-only]
	library_name: mlx
	pipeline_tag: text-generation
	language: [en]
	datasets: [nvidia/Nemotron-Image-Training-v3]
	inference: false
	---

	> [!TIP]
	> KV-cache quantization without any fork (recommended, 2026): upstream
	> llama.cpp/Ollama now cover this natively — use `-ctk q8_0 -ctv q8_0`
	> (~half KV memory, negligible quality loss: perplexity +0.002–0.05) or
	> `-ctk q4_0 -ctv q4_0` (~quarter memory, ≈7.6% perplexity increase). In
	> Ollama: `OLLAMA_KV_CACHE_TYPE=q8_0` with `OLLAMA_FLASH_ATTENTION=1`. Keep
	> K and V types symmetric to stay on the fast fused Flash-Attention path.
	> Since April 2026, mainline llama.cpp also applies Hadamard rotation to
	> KV activations ([PR #21038](https://github.com/ggml-org/llama.cpp/pull/21038)),
	> which greatly improves low-bit KV quality (opt-out:
	> `LLAMA_ATTN_ROT_DISABLE=1`).
	>
	> The RotorQuant/TurboQuant fork flow below is experimental/legacy: the
	> TurboQuant llama.cpp PR was closed without merging (June 2026) and the fork
	> is unmaintained relative to mainline. It is NOT required to use this model.
	<!-- kv-upstream-note -->

	# Nemotron-3-Nano-Omni-30B-A3B-Reasoning - TurboQuant MLX 3-bit

	MLX 3-bit quantization of the text tower of `Nemotron-3-Nano-Omni-30B-A3B-Reasoning` (`nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16`)
	with TurboQuant weight method. Apple Silicon native via `mlx-lm`.

	This variant covers the LLM backbone only. Vision (CRADIO v4-H) + audio (Parakeet-TDT-0.6B-v2)
	encoders are NOT included — MLX-VLM Nemotron-Omni model class is pending upstream support
	(no PR observed as of 2026-05-04). For multimodal inference, use the GGUF variants with
	`llama-mtmd-cli` instead.

	For the matched-KV stack — TurboQuant weights + TurboQuant KV-cache modifier —
	see [`majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit-TQ-KV`](https://huggingface.co/majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit-TQ-KV).
	For the runtime KV-cache modifier itself, see
	[`majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant`](https://huggingface.co/majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant).

	## Quickstart

	```python
	# Today (mlx-lm 0.31.x): the NemotronH_Nano_Omni_Reasoning_V3 model class
	# is not yet registered in mlx-lm. The cell below is the API shape that WILL
	# work once upstream lands the class (track ml-explore/mlx-lm#386).

	from mlx_lm import load, generate

	model, tokenizer = load("majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-3bit")

	prompt = tokenizer.apply_chat_template(
	[{"role": "user", "content": "Solve: 17 * 23"}],
	add_generation_prompt=True,
	enable_thinking=False, # set True to enable extended reasoning (default)
	)

	response = generate(
	model, tokenizer,
	prompt=prompt,
	max_tokens=512,
	sampler=lambda x: x.argmax(axis=-1), # or use mlx_lm.sample_utils.make_sampler(temp=0.6, top_p=0.95)
	)
	print(response)
	```

	> ⚠️ This variant covers the text tower only. For multimodal inference (vision + audio + video), use the GGUF variants with `llama-mtmd-cli` — see the GGUF cards in this family.

	## Modality matrix

	\| Modality \| Encoder \| Quantization in this variant \|
	\|---\|---\|---\|
	\| Text \| LLM backbone (Mamba-2 + Transformer hybrid Sparse MoE) \| per the variant suffix \|
	\| Image \| CRADIO v4-H \| BF16 (kept full-precision in every non-GGUF variant; GGUF uses mmproj-F16 split file) \|
	\| Audio \| Parakeet-TDT-0.6B-v2 \| BF16 (same rationale) \|
	\| Video \| Parakeet-TDT-0.6B-v2 + frame sampler \| BF16 (≤ 2 min, 256 frames @ 2 FPS) \|

	NVIDIA's official FP8 / NVFP4 recipe keeps both encoders + the cross-modal
	MLP projectors in BF16 to preserve multimodal accuracy. We follow that
	convention in every quantized variant we ship.

	## Runtime quirks

	### MLX-LM (text-only)

	This variant covers the LLM backbone only. Vision + audio encoders
	are NOT included — MLX-VLM Nemotron-Omni model class is
	pending upstream support (no PR observed as of 2026-05-04).

	Use the `mlx_lm.generate` API; `enable_thinking` is a runtime flag
	(see below).

	### Reasoning mode

	`enable_thinking` defaults to `True`. To disable extended reasoning
	(e.g., for latency-sensitive cases), pass `enable_thinking=False`
	to the chat template / generate call. No separate "no-think"
	variant card exists — this is a runtime flag, not a model variant.

	## Quant trade-off (MLX lane)

	\| Bits \| Approx size \| Use case \| Recommendation \|
	\|---\|---\|---\|---\|
	\| 2-bit \| ~8.1 GB \| Aggressive quantization \| Very low-RAM Macs \|
	\| 3-bit \| ~11 GB \| Lossy but small \| Low-RAM Macs \|
	\| 4-bit \| ~13 GB \| Balanced default \| Recommended for most Macs \|
	\| 5-bit \| ~16 GB \| Higher fidelity \| Quality-sensitive \|
	\| 6-bit \| ~19 GB \| Approaching FP16 quality \| High-fidelity \|
	\| 8-bit \| ~24 GB \| Near-lossless reference \| Fidelity-critical work \|

	(Current variant — 3bit — is bolded.)

	## Variants in this family

	(Showing 56 sibling variants under `majentik/nemotron3-nano-omni-30b-`. The current variant — `TurboQuant-MLX-3bit` — is bolded*.)

	\| Variant \| Runtime \| Approx size \| Use case \|
	\|---\|---\|---\|---\|
	\| [mmproj-F16](https://huggingface.co/majentik/nemotron3-nano-omni-30b-mmproj-f16) \| llama-mtmd-cli \| ~1-2 GB \| Multimodal projector (pair with any GGUF) \|
	\| [RotorQuant](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant) \| runtime modifier \| n/a \| KV-cache root (weight-agnostic) \|
	\| [RotorQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-IQ4_XS) \| llama.cpp \| ~26 GB \| Lossy 4-bit, low-RAM CPU/edge \|
	\| [RotorQuant-GGUF-MXFP4_MOE](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-MXFP4_MOE) \| llama.cpp \| ~30 GB \| MXFP4 MoE quant \|
	\| [RotorQuant-GGUF-Q2_K](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q2_K) \| llama.cpp \| ~18 GB \| Lossy, low-RAM CPU/edge \|
	\| [RotorQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q3_K_M) \| llama.cpp \| ~23 GB \| Smaller 3-bit, CPU-friendly \|
	\| [RotorQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q4_K_M) \| llama.cpp \| ~33 GB \| Balanced default \|
	\| [RotorQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q5_K_M) \| llama.cpp \| ~40 GB \| Higher fidelity, more RAM \|
	\| [RotorQuant-GGUF-Q8_0](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-Q8_0) \| llama.cpp \| ~63 GB \| Near-lossless reference \|
	\| [RotorQuant-GGUF-IQ4_XS-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-iq4_xs-rq-kv) \| llama.cpp \| ~26 GB \| IQ4_XS + RotorQuant KV \|
	\| [RotorQuant-GGUF-MXFP4_MOE-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-mxfp4_moe-rq-kv) \| llama.cpp \| ~30 GB \| MXFP4 MoE + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q2_K-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q2_k-rq-kv) \| llama.cpp \| ~18 GB \| Q2_K + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q3_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q3_k_m-rq-kv) \| llama.cpp \| ~23 GB \| Q3_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q4_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q4_k_m-rq-kv) \| llama.cpp \| ~33 GB \| Q4_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q5_K_M-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q5_k_m-rq-kv) \| llama.cpp \| ~40 GB \| Q5_K_M + RotorQuant KV \|
	\| [RotorQuant-GGUF-Q8_0-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-gguf-q8_0-rq-kv) \| llama.cpp \| ~63 GB \| Q8_0 + RotorQuant KV \|
	\| [RotorQuant-MLX-2bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-2bit) \| mlx-lm \| ~9.6 GB \| Apple Silicon, smallest \|
	\| [RotorQuant-MLX-2bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-2bit-rq-kv) \| mlx-lm \| ~9.6 GB \| 2-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-3bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-3bit) \| mlx-lm \| ~14 GB \| Apple Silicon, small \|
	\| [RotorQuant-MLX-3bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-3bit-rq-kv) \| mlx-lm \| ~14 GB \| 3-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-4bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-4bit) \| mlx-lm \| ~19 GB \| Apple Silicon balanced \|
	\| [RotorQuant-MLX-4bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-4bit-rq-kv) \| mlx-lm \| ~19 GB \| 4-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-5bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-5bit) \| mlx-lm \| ~23 GB \| Apple Silicon, higher fidelity \|
	\| [RotorQuant-MLX-5bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-5bit-rq-kv) \| mlx-lm \| ~23 GB \| 5-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-6bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-6bit) \| mlx-lm \| ~27 GB \| Apple Silicon, near-lossless \|
	\| [RotorQuant-MLX-6bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-6bit-rq-kv) \| mlx-lm \| ~27 GB \| 6-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-8bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-8bit) \| mlx-lm \| ~35 GB \| Apple Silicon reference \|
	\| [RotorQuant-MLX-8bit-RQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-8bit-rq-kv) \| mlx-lm \| ~35 GB \| 8-bit + RotorQuant KV \|
	\| [RotorQuant-MLX-MXFP4](https://huggingface.co/majentik/nemotron3-nano-omni-30b-rotorquant-mlx-mxfp4) \| mlx-lm \| ~19 GB \| Apple Silicon MXFP4 \|
	\| [TurboQuant](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant) \| runtime modifier \| n/a \| KV-cache root (weight-agnostic) \|
	\| [TurboQuant-GGUF-IQ4_XS](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-IQ4_XS) \| llama.cpp \| ~26 GB \| Lossy 4-bit, low-RAM CPU/edge \|
	\| [TurboQuant-GGUF-MXFP4_MOE](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-MXFP4_MOE) \| llama.cpp \| ~30 GB \| MXFP4 MoE quant \|
	\| [TurboQuant-GGUF-Q2_K](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q2_K) \| llama.cpp \| ~18 GB \| Lossy, low-RAM CPU/edge \|
	\| [TurboQuant-GGUF-Q3_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q3_K_M) \| llama.cpp \| ~23 GB \| Smaller 3-bit, CPU-friendly \|
	\| [TurboQuant-GGUF-Q4_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q4_K_M) \| llama.cpp \| ~33 GB \| Balanced default \|
	\| [TurboQuant-GGUF-Q5_K_M](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q5_K_M) \| llama.cpp \| ~40 GB \| Higher fidelity, more RAM \|
	\| [TurboQuant-GGUF-Q8_0](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-Q8_0) \| llama.cpp \| ~63 GB \| Near-lossless reference \|
	\| [TurboQuant-GGUF-IQ4_XS-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-iq4_xs-tq-kv) \| llama.cpp \| ~26 GB \| IQ4_XS + TurboQuant KV \|
	\| [TurboQuant-GGUF-MXFP4_MOE-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-mxfp4_moe-tq-kv) \| llama.cpp \| ~30 GB \| MXFP4 MoE + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q2_K-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q2_k-tq-kv) \| llama.cpp \| ~18 GB \| Q2_K + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q3_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q3_k_m-tq-kv) \| llama.cpp \| ~23 GB \| Q3_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q4_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q4_k_m-tq-kv) \| llama.cpp \| ~33 GB \| Q4_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q5_K_M-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q5_k_m-tq-kv) \| llama.cpp \| ~40 GB \| Q5_K_M + TurboQuant KV \|
	\| [TurboQuant-GGUF-Q8_0-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-gguf-q8_0-tq-kv) \| llama.cpp \| ~63 GB \| Q8_0 + TurboQuant KV \|
	\| [TurboQuant-MLX-2bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-2bit) \| mlx-lm \| ~9.6 GB \| Apple Silicon, smallest \|
	\| [TurboQuant-MLX-2bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-2bit-tq-kv) \| mlx-lm \| ~9.6 GB \| 2-bit + TurboQuant KV \|
	\| TurboQuant-MLX-3bit \| mlx-lm \| ~14 GB \| Apple Silicon, small \|
	\| [TurboQuant-MLX-3bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-3bit-tq-kv) \| mlx-lm \| ~14 GB \| 3-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-4bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-4bit) \| mlx-lm \| ~19 GB \| Apple Silicon balanced \|
	\| [TurboQuant-MLX-4bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-4bit-tq-kv) \| mlx-lm \| ~19 GB \| 4-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-5bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-5bit) \| mlx-lm \| ~23 GB \| Apple Silicon, higher fidelity \|
	\| [TurboQuant-MLX-5bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-5bit-tq-kv) \| mlx-lm \| ~23 GB \| 5-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-6bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-6bit) \| mlx-lm \| ~27 GB \| Apple Silicon, near-lossless \|
	\| [TurboQuant-MLX-6bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-6bit-tq-kv) \| mlx-lm \| ~27 GB \| 6-bit + TurboQuant KV \|
	\| [TurboQuant-MLX-8bit](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-8bit) \| mlx-lm \| ~35 GB \| Apple Silicon reference \|
	\| [TurboQuant-MLX-8bit-TQ-KV](https://huggingface.co/majentik/nemotron3-nano-omni-30b-turboquant-mlx-8bit-tq-kv) \| mlx-lm \| ~35 GB \| 8-bit + TurboQuant KV \|