Best AI Models for NVIDIA Jetson Orin NX 16GB
16 GB total — ~13 GB usable as VRAM
Embedded edge module (16 GB, 102 GB/s) — sized for robotics/IoT; small-to-mid models only.
16 GB is a comfortable mid-range tier for local AI. Most 7B–13B models run smoothly at good quantization levels, and smaller models can run at near-full precision.
This memory tier strikes a nice balance between price and capability. Popular 7B models like Llama 3 8B, Mistral 7B, and Qwen 2.5 7B all run very well at Q4_K_M quantization with fast inference and reasonable context windows. You can also fit some larger 13B models at Q3–Q4, though you'll want to keep context lengths modest. Small models like Phi 3 Mini (3.8B) practically fly at Q8 or even FP16 quality.
Runs Well
- 7B models at Q4–Q6 quality with good speed
- Small models (3B–4B) at Q8 or FP16
- 9B models (Gemma 2 9B) at Q4_K_M
Challenging
- 13B–14B models need Q3 or lower
- 30B+ models do not fit in VRAM
- Long context (>8K tokens) with larger models
What LLMs Can NVIDIA Jetson Orin NX 16GB Run?
206 models · 24 excellent · 36 good
Showing compatibility for NVIDIA Jetson Orin NX 16GB
| Model | Quant | VRAM | Speed | Context | Status | Grade |
|---|---|---|---|---|---|---|
Q4_K_M·85.4 t/s tok/s·128K ctx·RUNS GREAT | Q4_K_M | 5.5 GB | 85.4 t/s | 128K | RUNS GREAT | S88 |
Q4_K_M·40.6 t/s tok/s·131K ctx·RUNS WELL | Q4_K_M | 7.7 GB | 40.6 t/s | 131K | RUNS WELL | A75 |
Q4_K_M·65.1 t/s tok/s·131K ctx·RUNS WELL | Q4_K_M | 5.3 GB | 65.1 t/s | 131K | RUNS WELL | A84 |
Q4_K_M·76.1 t/s tok/s·33K ctx·RUNS GREAT | Q4_K_M | 2.4 GB | 76.1 t/s | 33K | RUNS GREAT | S87 |
Q4_K_M·10.5 t/s tok/s·262K ctx·DECENT | Q4_K_M | 6.4 GB | 10.5 t/s | 262K | DECENT | B49 |
Q4_K_M·8.1 t/s tok/s·262K ctx·DECENT | Q4_K_M | 8.2 GB | 8.1 t/s | 262K | DECENT | B45 |
Q4_K_M·40.6 t/s tok/s·131K ctx·RUNS WELL | Q4_K_M | 7.7 GB | 40.6 t/s | 131K | RUNS WELL | A75 |
Q4_K_M·73.1 t/s tok/s·8K ctx·RUNS GREAT | Q4_K_M | 2.5 GB | 73.1 t/s | 8K | RUNS GREAT | S86 |
Q4_K_M·10.7 t/s tok/s·262K ctx·DECENT | Q4_K_M | 6.2 GB | 10.7 t/s | 262K | DECENT | B49 |
Q4_K_M·10.5 t/s tok/s·262K ctx·DECENT | Q4_K_M | 6.4 GB | 10.5 t/s | 262K | DECENT | B49 |
Q4_K_M·32.6 t/s tok/s·131K ctx·RUNS WELL | Q4_K_M | 2.0 GB | 32.6 t/s | 131K | RUNS WELL | A72 |
Q4_K_M·12.5 t/s tok/s·131K ctx·DECENT | Q4_K_M | 5.3 GB | 12.5 t/s | 131K | DECENT | B51 |
Q4_K_M·76.1 t/s tok/s·8K ctx·RUNS GREAT | Q4_K_M | 2.4 GB | 76.1 t/s | 8K | RUNS GREAT | S87 |
Q4_K_M·11.3 t/s tok/s·DECENT | Q4_K_M | 5.9 GB | 11.3 t/s | — | DECENT | B50 |
Q2_K·34.9 t/s tok/s·262K ctx·DECENT | Q2_K | 11.6 GB | 34.9 t/s | 262K | DECENT | B60 |
Q4_K_M·20.4 t/s tok/s·262K ctx·DECENT | Q4_K_M | 3.3 GB | 20.4 t/s | 262K | DECENT | B60 |
NVIDIA Jetson Orin NX 16GB Specifications
- Brand
- NVIDIA
- Chip
- Orin NX
- Type
- AI Box
- Unified Memory
- 16 GB
- Memory Bandwidth
- 102.4 GB/s
- GPU Cores
- 1024
- CPU Cores
- 8
- Neural Engine
- 157.0 TOPS
- Architecture
- Ampere
- Memory Type
- LPDDR5
- TDP
- 10-40 W
- MSRP
- $599
- Release Date
- 2023-02-01
Get Started
Prompt Processing
Estimated for LLaDA2.0 Mini, the compute-bound phase that reads your prompt before the first reply token appears.
Short chat
217 ms
512 tok prompt
Long chat
1.7 s
4,096 tok prompt
Document / codebase
13.9 s
32,768 tok prompt
Prefill is compute-bound and a different number from the decode tok/s shown elsewhere on this page — how prompt processing works →
Performance figures are estimates calibrated as of 2026-09-21 — see calibration basis →
Devices to Consider
Similar devices and upgrades with more memory or higher bandwidth
Frequently Asked Questions
- Can NVIDIA Jetson Orin NX 16GB run LFM2.5 8B A1B?
Yes, the NVIDIA Jetson Orin NX 16GB with 16 GB unified memory can run LFM2.5 8B A1B, Mellum2 12B A2.5B Instruct, Ling 3.0 Tiny, and 1978 other models. 466 models achieve excellent performance, and 377 run at good quality. Apple Silicon's unified memory architecture lets the GPU access the full memory pool without copying data, making it efficient for AI workloads.
- How much memory is available for AI on NVIDIA Jetson Orin NX 16GB?
The NVIDIA Jetson Orin NX 16GB has 16 GB unified memory. After macOS reserves ~3.5 GB for the operating system, approximately 12.5 GB is available for AI models. Unlike discrete GPUs where VRAM is separate from system RAM, Apple Silicon shares one memory pool between the CPU and GPU — this means no data copying overhead, but you share memory with macOS and open apps.
- Is NVIDIA Jetson Orin NX 16GB good for AI?
With 16 GB unified memory and 102.4 GB/s bandwidth, the NVIDIA Jetson Orin NX 16GB is good for running local AI models. It supports 843 models at good quality or better. It's a capable entry point for 7B models. Apple Silicon's Metal acceleration and unified memory make it surprisingly efficient despite the modest memory.
- What's the best model for NVIDIA Jetson Orin NX 16GB?
The top-rated models for the NVIDIA Jetson Orin NX 16GB are LFM2.5 8B A1B, Mellum2 12B A2.5B Instruct, Ling 3.0 Tiny. At this memory level, 7B models at Q4_K_M give you the best experience — fast responses and solid quality for chat and coding assistance.
- How fast is NVIDIA Jetson Orin NX 16GB for AI inference?
With 102.4 GB/s memory bandwidth, the NVIDIA Jetson Orin NX 16GB achieves approximately 16 tok/s on a 7B model at Q4_K_M — that's functional for interactive use. A 14B model runs at ~8 tok/s. Apple Silicon achieves high efficiency (~70%) thanks to unified memory — there's no PCIe bottleneck between CPU and GPU.
tok/s = (102.4 GB/s ÷ model GB) × efficiency
Apple Silicon achieves ~70% bandwidth efficiency thanks to unified memory and Metal acceleration.
Estimated speed on NVIDIA Jetson Orin NX 16GB
~85 tok/s~41 tok/s~65 tok/s~76 tok/sReal-world results typically within ±20%.
- Can I run AI offline on NVIDIA Jetson Orin NX 16GB?
Yes — once you download a model, it runs entirely on the NVIDIA Jetson Orin NX 16GB without internet. Applications like Ollama and LM Studio make it straightforward to download, manage, and run models locally. All your conversations stay private on your device with zero data sent to external servers. This is one of the key advantages of local AI: complete privacy, no API costs, and no rate limits.
- Anything to watch out for with NVIDIA Jetson Orin NX 16GB?
Embedded edge module (16 GB, 102 GB/s) — sized for robotics/IoT; small-to-mid models only.