Featured Models

Latest releases with day-0 support on Jetson

Nemotron 3.5 Bolt E4B 4GB

LLM

NVIDIA's compact conversational model for local assistants, multi-turn tool use, and interactive characters. Run text chat on Jetson with llama.cpp.

Nemotron 3.5 Bolt E4B 4GB

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

d1-3B

VLM

Liquid AI's zero-shot classification model for text and images, with a guide to running it on Jetson using PyTorch.

Qwen3.8 Flash Next

VLM

Qwen's vision-language MoE model for reasoning, coding, and tool use on Jetson Thor, with NVFP4 quantization and MTP speculative decoding

Qwen3.8 Flash Next

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

All Models

Browse by family cards or switch to a sortable table.

56 models

Agentic & decision

Liquid AI d1

d1-3B New

Classification / decision

d1-omni-600M

Classification / decision

Physical AI & action

FLUX.3 Action

FLUX 3 Action

Action generation / VLA

General & multimodal

Qwen 3.8

Qwen3.8 Flash Next New

Text generation

Qwen3.8 Flash Next

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.8 27B

Text generation

Qwen3.8 27B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

NVIDIA models

NVIDIA Nemotron

Nemotron 3.5 Bolt E4B 4GB New

Text generation

Nemotron 3.5 Bolt E4B 4GB

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron 3.5 Lightning

Text generation

Nemotron 3.5 Lightning

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron3 Nano 4B

Text generation

Nemotron3 Nano 4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron3 Nano 30B-A3B

Text generation

Nemotron3 Nano 30B-A3B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron Nano 9B v2

Text generation

Nemotron Nano 9B v2

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron 3 Super 120B-A12B

Text generation

Nemotron 3 Super 120B-A12B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron Nano 12B VL

Text generation

Nemotron Nano 12B VL

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Nemotron 3 Nano Omni

Text generation

Nemotron 3 Nano Omni

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Agentic & decision

Meta Muse Glimmer

Muse Glimmer 30B

Text generation

Muse Glimmer 30B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Qwen 3.6

Qwen3.6 35B-A3B (MoE)

Text generation

Qwen3.6 35B-A3B (MoE)

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.6 27B

Text generation

Qwen3.6 27B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Gemma 4

Gemma 4 E2B

Text generation

Gemma 4 E2B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 4 12B

Text generation

Gemma 4 12B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 4 E4B

Text generation

Gemma 4 E4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 4 26B-A4B

Text generation

Gemma 4 26B-A4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 4 31B

Text generation

Gemma 4 31B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

DiffusionGemma 26B-A4B

Text generation

DiffusionGemma 26B-A4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

NVIDIA models

NVIDIA Cosmos

Cosmos Reason 1 7B

Text generation

Cosmos Reason 1 7B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Cosmos Reason 2 2B

Text generation

Cosmos3 Edge

Action generation / VLA

Cosmos3 Edge

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Cosmos Reason 2 8B

Text generation

Cosmos3 Nano

Text generation

General & multimodal

Qwen 3.5

Qwen3.5 35B-A3B (MoE)

Text generation

Qwen3.5 35B-A3B (MoE)

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.5 27B

Text generation

Qwen3.5 27B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.5 9B

Text generation

Qwen3.5 9B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.5 4B

Text generation

Qwen3.5 4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3.5 0.8B

Text generation

Qwen3.5 0.8B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

OpenAI GPT OSS

GPT OSS 20B

Text generation

GPT OSS 20B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

GPT OSS 120B

Text generation

GPT OSS 120B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Qwen 3

Qwen3 4B

Text generation

Qwen3 4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3 8B

Text generation

Qwen3 8B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3 30B-A3B (MoE)

Text generation

Qwen3 30B-A3B (MoE)

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3 32B

Text generation

Qwen3 32B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3 VL 4B

Text generation

Qwen3 VL 4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Qwen3 VL 8B

Text generation

Qwen3 VL 8B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Ministral 3

Ministral 3 3B Instruct

Text generation

Ministral 3 3B Instruct

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Ministral 3 8B Instruct

Text generation

Ministral 3 8B Instruct

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Ministral 3 14B Instruct

Text generation

Ministral 3 14B Instruct

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Ministral 3 3B Reasoning

Text generation

Ministral 3 3B Reasoning

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Ministral 3 8B Reasoning

Text generation

Ministral 3 8B Reasoning

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Ministral 3 14B Reasoning

Text generation

Ministral 3 14B Reasoning

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Llama 3

Llama 3.2 3B

Text generation

Llama 3.2 3B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Llama 3.1 8B

Text generation

Llama 3.1 8B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Llama 3.1 70B

Text generation

Llama 3.1 70B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

General & multimodal

Gemma 3

FunctionGemma

Text generation

FunctionGemma

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 3 270M

Text generation

Gemma 3 270M

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 3 1B

Text generation

Gemma 3 1B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 3 4B

Text generation

Gemma 3 4B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 3 12B

Text generation

Gemma 3 12B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Gemma 3 27B

Text generation

Gemma 3 27B

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

Agentic & decision

MiniMax M2.7

Text generation

MiniMax M2.7

Quick Start Runner

Inference Engine

Loading command...

Commands are auto-generated based on your configuration settings.

All models: filter by model name in the Model column; sort Model, Family, or Output; use Modules and Inference engines checkboxes to filter further.
Filter by model name		Modules					Inference engines				Actions
Filter by model name		T5000	T4000	AGX Orin 64GB	Orin NX 16GB	Orin Nano 8GB	vLLM	Ollama	llama.cpp	TensorRT Edge-LLM	Actions
Nemotron 3.5 Bolt E4B 4GB New	NVIDIA Nemotron	✓	✓	✓	✓	✓	—	—	✓	—	Nemotron 3.5 Bolt E4B 4GB Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
d1-3B New	Liquid AI d1	—	—	—	—	—	—	—	—	—
d1-omni-600M New	Liquid AI d1	—	—	—	—	—	—	—	—	—
FLUX 3 Action New	Black Forest Labs	—	—	—	—	—	—	—	—	—
Qwen3.8 Flash Next New	Alibaba Qwen3.8	✓	—	—	—	—	✓	—	—	—	Qwen3.8 Flash Next Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.8 27B New	Alibaba Qwen3.8	✓	✓	✓	—	—	—	—	✓	—	Qwen3.8 27B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron 3.5 Lightning New	NVIDIA Nemotron	✓	✓	✓	—	—	✓	—	✓	✓	Nemotron 3.5 Lightning Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Muse Glimmer 30B New	Meta Muse	✓	✓	✓	—	—	—	—	✓	—	Muse Glimmer 30B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron3 Nano 4B New	NVIDIA Nemotron	✓	✓	✓	✓	✓	—	—	✓	—	Nemotron3 Nano 4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
FunctionGemma New	Google Gemma3	✓	✓	✓	✓	✓	—	—	✓	—	FunctionGemma Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Cosmos Reason 1 7B New	NVIDIA Cosmos	✓	✓	✓	✓	✓	✓	—	—	—	Cosmos Reason 1 7B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 3 270M New	Google Gemma3	✓	✓	✓	✓	✓	✓	✓	—	—	Gemma 3 270M Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 4 E2B New	Google Gemma4	✓	✓	✓	✓	✓	✓	—	✓	—	Gemma 4 E2B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
GPT OSS 20B New	OpenAI GPT OSS	✓	✓	✓	—	—	✓	—	—	—	GPT OSS 20B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Llama 3.2 3B New	Meta Llama 3	✓	✓	✓	✓	✓	✓	✓	—	—	Llama 3.2 3B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
MiniMax M2.7 New	MiniMax M2.7	✓	—	—	—	—	—	—	✓	—	MiniMax M2.7 Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 3B Instruct New	Mistral AI Ministral 3	✓	✓	✓	✓	✓	✓	✓	—	—	Ministral 3 3B Instruct Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron3 Nano 30B-A3B New	NVIDIA Nemotron	✓	✓	✓	✓	✓	✓	✓	—	—	Nemotron3 Nano 30B-A3B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 4B New	Alibaba Qwen3	✓	✓	✓	✓	✓	✓	—	—	—	Qwen3 4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.5 35B-A3B (MoE) New	Alibaba Qwen3.5	✓	✓	✓	—	—	✓	—	—	—	Qwen3.5 35B-A3B (MoE) Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.6 35B-A3B (MoE) New	Alibaba Qwen3.6	✓	✓	✓	—	—	✓	—	—	—	Qwen3.6 35B-A3B (MoE) Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 3 1B New	Google Gemma3	✓	✓	✓	✓	✓	✓	✓	—	—	Gemma 3 1B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 4 12B New	Google Gemma4	✓	✓	✓	—	—	✓	—	✓	—	Gemma 4 12B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 4 E4B New	Google Gemma4	✓	✓	✓	✓	✓	✓	—	✓	—	Gemma 4 E4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
GPT OSS 120B New	OpenAI GPT OSS	✓	✓	—	—	—	✓	—	—	—	GPT OSS 120B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Llama 3.1 8B New	Meta Llama 3	✓	✓	✓	✓	✓	✓	✓	—	—	Llama 3.1 8B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 8B Instruct New	Mistral AI Ministral 3	✓	✓	✓	✓	✓	✓	✓	—	—	Ministral 3 8B Instruct Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron Nano 9B v2 New	NVIDIA Nemotron	✓	✓	—	—	—	✓	—	—	—	Nemotron Nano 9B v2 Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.5 27B New	Alibaba Qwen3.5	✓	✓	✓	—	—	✓	—	—	—	Qwen3.5 27B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.6 27B New	Alibaba Qwen3.6	✓	✓	✓	—	—	✓	—	—	—	Qwen3.6 27B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 8B New	Alibaba Qwen3	✓	✓	✓	✓	—	✓	—	—	—	Qwen3 8B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Cosmos Reason 2 2B New	NVIDIA Cosmos	✓	✓	✓	✓	✓	✓	—	✓	—
Gemma 3 4B New	Google Gemma3	✓	✓	✓	✓	✓	✓	✓	—	—	Gemma 3 4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 4 26B-A4B New	Google Gemma4	✓	✓	✓	—	—	✓	—	✓	—	Gemma 4 26B-A4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Llama 3.1 70B New	Meta Llama 3	✓	✓	✓	✓	✓	✓	✓	—	—	Llama 3.1 70B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 14B Instruct New	Mistral AI Ministral 3	✓	✓	✓	✓	✓	✓	✓	—	—	Ministral 3 14B Instruct Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron 3 Super 120B-A12B New	NVIDIA Nemotron	✓	—	—	—	—	✓	—	✓	—	Nemotron 3 Super 120B-A12B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron Nano 12B VL New	NVIDIA Nemotron	✓	✓	—	—	—	✓	—	—	—	Nemotron Nano 12B VL Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 30B-A3B (MoE) New	Alibaba Qwen3	✓	✓	✓	—	—	✓	—	—	—	Qwen3 30B-A3B (MoE) Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.5 9B New	Alibaba Qwen3.5	✓	✓	✓	✓	—	✓	—	—	—	Qwen3.5 9B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Cosmos3 Edge New	NVIDIA Cosmos	✓	✓	✓	—	—	✓	—	—	—	Cosmos3 Edge Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Gemma 3 12B New	Google Gemma3	✓	✓	✓	✓	✓	✓	✓	—	—	Gemma 3 12B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Cosmos Reason 2 8B New	NVIDIA Cosmos	✓	✓	✓	✓	✓	✓	—	✓	—
Gemma 4 31B New	Google Gemma4	✓	✓	✓	—	—	✓	—	✓	—	Gemma 4 31B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 3B Reasoning New	Mistral AI Ministral 3	✓	✓	✓	✓	✓	✓	—	—	—	Ministral 3 3B Reasoning Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 32B New	Alibaba Qwen3	✓	✓	—	—	—	✓	—	—	—	Qwen3 32B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Nemotron 3 Nano Omni New	NVIDIA Nemotron	✓	✓	✓	—	—	✓	✓	✓	—	Nemotron 3 Nano Omni Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.5 4B New	Alibaba Qwen3.5	✓	✓	✓	✓	✓	✓	—	—	—	Qwen3.5 4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
DiffusionGemma 26B-A4B New	Google Gemma4	✓	✓	✓	—	—	✓	—	—	—	DiffusionGemma 26B-A4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Cosmos3 Nano New	NVIDIA Cosmos	✓	✓	—	—	—	✓	—	—	—
Gemma 3 27B New	Google Gemma3	✓	✓	✓	✓	✓	✓	✓	—	—	Gemma 3 27B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 8B Reasoning New	Mistral AI Ministral 3	✓	✓	✓	✓	—	✓	—	—	—	Ministral 3 8B Reasoning Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3.5 0.8B New	Alibaba Qwen3.5	✓	✓	✓	✓	✓	✓	—	—	—	Qwen3.5 0.8B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 VL 4B New	Alibaba Qwen3	✓	✓	✓	✓	✓	✓	—	—	—	Qwen3 VL 4B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Ministral 3 14B Reasoning New	Mistral AI Ministral 3	✓	✓	✓	—	—	✓	—	—	—	Ministral 3 14B Reasoning Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)
Qwen3 VL 8B New	Alibaba Qwen3	✓	✓	✓	✓	—	✓	—	—	—	Qwen3 VL 8B Quick Start Runner Inference Engine `Loading command...` Commands are auto-generated based on your configuration settings. Advanced configuration Authentication Hugging Face Token * This model requires a Hugging Face access token. The token is inserted into the command and never stored. vLLM Configuration Configure vLLM server parameters. Leave empty to use defaults. Port (--port) Max Model Length (--max-model-len) Maximum context length the model can handle GPU Memory Utilization (--gpu-memory-utilization) Fraction of GPU memory to use (0.1 - 1.0)

Performance Comparison

Benchmarks on Jetson across supported inference runtimes

Platform

Concurrency

* ISL/OSL for all benchmarks: 2048/128

* Unless otherwise specified, all models utilize W4A16 quantization for Orin and NVFP4 for Thor.

* NVFP4 and MXFP4 require Blackwell FP4 tensor cores and are not available on Orin (Ampere).

* Edge-LLM benchmark values are generation throughput from the official TensorRT Edge-LLM performance benchmarks on Jetson AGX Thor Developer Kit.

* Cosmos3 Edge (Reasoner) (BF16): streaming chat harness, ISL 1705 / OSL 128, per-request prompt-cache salting, median of 3 - not the aiperf ISL-2048 protocol used for other rows.

Featured Models

Nemotron 3.5 Bolt E4B 4GB

Nemotron 3.5 Bolt E4B 4GB

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

d1-3B

Qwen3.8 Flash Next

Qwen3.8 Flash Next

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

All Models

Liquid AI d1

FLUX.3 Action

Qwen 3.8

Qwen3.8 Flash Next

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Qwen3.8 27B

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

NVIDIA Nemotron

Nemotron 3.5 Bolt E4B 4GB

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron 3.5 Lightning

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron3 Nano 4B

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron3 Nano 30B-A3B

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron Nano 9B v2

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron 3 Super 120B-A12B

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron Nano 12B VL

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Nemotron 3 Nano Omni

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Meta Muse Glimmer

Muse Glimmer 30B

Inference Engine

Advanced configuration

Authentication

vLLM Configuration

Qwen 3.6

Qwen3.6 35B-A3B (MoE)

Inference Engine

Advanced configuration

Authentication