Instructions to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Notebooks
Google Colab
Kaggle
Local Apps Settings

How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Use Docker

docker model run hf.co/YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with Ollama:
```
ollama run hf.co/YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
```
Unsloth Desktop

How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with Pi:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Configure the model in Pi

# Install Pi:
npm install -g @earendil-works/pi-coding-agent
# Add to ~/.pi/agent/models.json:
{
  "providers": {
    "llama-cpp": {
      "baseUrl": "http://localhost:8080/v1",
      "api": "openai-completions",
      "apiKey": "none",
      "models": [
        {
          "id": "YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M"
        }
      ]
    }
  }
}

Run Pi

# Start Pi in your project directory:
pi

Docker Model Runner
How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with Docker Model Runner:
```
docker model run hf.co/YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M
```

Lemonade

How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.Nemotron-3-Puzzle-75B-A9B-GGUF-Q4_K_M

List all available models

lemonade list

Hermes Agent

How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with Hermes Agent:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

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 YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

Run Hermes

hermes

Atomic Chat

OpenClaw

How to use YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF with OpenClaw:

Start the llama.cpp server

# Install llama.cpp:
brew install llama.cpp
# Start a local OpenAI-compatible server:
llama serve -hf YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M

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 "YanissAmz/Nemotron-3-Puzzle-75B-A9B-GGUF:Q4_K_M" \
  --custom-provider-id llama-cpp \
  --custom-compatibility openai \
  --custom-text-input \
  --accept-risk \
  --skip-health

Run OpenClaw

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

Nemotron-3-Puzzle-75B-A9B — GGUF

First GGUF release of NVIDIA's Nemotron-3-Puzzle-75B-A9B (hybrid mamba2/attention/latent-MoE, 75B total / 9B active, 262k context, MTP draft head).

Converted from the official FP8 checkpoint (weight scales absorbed at conversion — no double quantization), then quantized from the Q8_0 master with an importance matrix.

Files

file	size	note
Puzzle-75B-A9B-Q8_0-0000X-of-00002.gguf	77.7 GiB (2 shards)	master, near-lossless — point llama.cpp at shard 00001, the rest loads automatically
Puzzle-75B-A9B-Q4_K_M-0000X-of-00002.gguf	48.1 GiB (2 shards)	reference k-quant, fastest decode
Puzzle-75B-A9B-NVFP4.gguf	45.0 GiB	experts NVFP4, everything else Q8_0
Puzzle-75B-A9B-UD-IQ4-XL.gguf	41.6 GiB	experts IQ4_XS; attn Q8_0, ssm/shexp Q6_K, ffn_latent Q8_0
puzzle-imatrix.gguf	0.2 GiB	reusable imatrix (calibration_datav3)

Requirements

Not yet supported by mainline llama.cpp — needs per-layer heterogeneous MoE arrays and the 2-sub-block MTP head. Use the puzzle-port branch until the PR is merged: [PR_LINK]

Measured (Strix Halo 128GB unified, Radeon 8060S, `-ngl 99`; PPL = wikitext-2 test, 24 chunks)

quant	PPL	decode t/s	prefill t/s	backend
Q8_0	5.325	10.2	189	Vulkan
Q4_K_M	5.404	19.9	238	ROCm
UD-IQ4-XL	5.377	17.7	211	ROCm
NVFP4	5.383	16.6	243	ROCm

All three 4-bit variants sit within noise of each other on PPL (±0.08); pick by speed/size trade-off.

⚠️ On Strix Halo (gfx1151) use the ROCm/HIP backend for the 4-bit quants: Vulkan decode collapses to ~2.7 t/s on this model's MoE (mul_mat_id slow path). Q8_0 exceeds the ROCm allocation limit → run it on Vulkan.

MTP speculative decoding (--spec-type draft-mtp) loads and drafts correctly, but is currently slower than plain decoding (~13 vs 16.6 t/s): llama.cpp cannot yet roll back mamba2 recurrent states, which throttles draft attempts. Leave it off for now.

Notes

Reasoning model: llama-server parses the thinking channel natively.
AI-assisted work; everything reviewed and validated end-to-end on my hardware.

Downloads last month: 1,330

GGUF

Model size

78B params

Architecture

nemotron_h_moe

Hardware compatibility

4-bit

8-bit

View +2 variants

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Nemotron-3-Puzzle-75B-A9B — GGUF

Files

Requirements

Measured (Strix Halo 128GB unified, Radeon 8060S, -ngl 99; PPL = wikitext-2 test, 24 chunks)

Notes

Measured (Strix Halo 128GB unified, Radeon 8060S, `-ngl 99`; PPL = wikitext-2 test, 24 chunks)