ASUS ROG Zephyrus G16 Intel Core Ultra 9 185H + NVIDIA GeForce RTX 4090 Performance Overview
The ASUS ROG Zephyrus G16, equipped with the Intel Core Ultra 9 185H and NVIDIA GeForce RTX 4090 Laptop GPU, provides robust AI performance across a variety of tasks. With 32 GB of RAM, it performs exceptionally in demanding applications such as Style Transfer, Pose Estimation, and Depth Estimation. The RTX 4090 is ideal for GPU-driven AI workloads, while the Intel Core Ultra 9 185H CPU and Intel AI Boost NPU add flexibility for CPU and NPU-optimized tasks.
Hardware Specs
- CPU: Intel Core Ultra 9 185H, 16 cores (2.50 GHz)
- RAM: 32 GB
- GPU: NVIDIA GeForce RTX 4090 Laptop GPU
- NPU: Intel AI Boost
Estimated Total Price (2024): $2,564.99 (open box at Best Buy)
Price-to-Performance: 877 per $100
The ASUS ROG Zephyrus G16 achieves a price-to-performance ratio of 877 per $100, based on its overall single precision GPU benchmark score of 22,486 with the NVIDIA GeForce RTX 4090.
Price-to-Performance Calculation:
The price-to-performance ratio is calculated by dividing the single precision GPU benchmark score by the system price, then multiplying by 100.
Calculation: (22486 / 2564.99) × 100 ≈ 877 per $100
DirectML GPU Benchmarks – NVIDIA GeForce RTX 4090
This benchmark highlights the NVIDIA GeForce RTX 4090 Laptop GPU, showcasing its powerful AI processing capabilities, especially in single and half precision tasks that are critical for demanding AI workloads.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 22,486
- Half Precision: 33,742
- Quantized: 17,513
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 274,854 | 353.3 | 4. Depth Estimation | 53,354 | 411.1 |
| 2. Pose Estimation | 147,014 | 171.5 | 5. Image Segmentation | 14,333 | 232.3 |
| 3. Face Detection | 29,595 | 351.7 | 6. Object Detection | 10,798 | 856.5 |
ONNX CPU Benchmarks – Intel Core Ultra 9 185H
This ONNX-based benchmark measures the Intel Core Ultra 9 185H CPU’s AI performance, which is strong in single and quantized precision tasks, with particularly notable results in Style Transfer and Depth Estimation. However, the GPU significantly outperforms the CPU in higher-demand AI applications.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 3,076
- Half Precision: 1,342
- Quantized: 3,875
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 14,709 | 18.9 | 4. Face Detection | 3,679 | 43.7 |
| 2. Depth Estimation | 6,919 | 53.3 | 5. Object Detection | 2,007 | 159.2 |
| 3. Pose Estimation | 6,248 | 7.29 | 6. Image Segmentation | 1,351 | 21.9 |
OpenVINO CPU Benchmarks – Intel Core Ultra 9 185H
This benchmark uses the OpenVINO Backend to evaluate the CPU’s AI performance. While the Intel Core Ultra 9 185H performs reasonably in quantized tasks, it is notably outpaced by the dedicated RTX 4090 GPU for complex AI workloads.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 3,009
- Half Precision: 2,992
- Quantized: 7,794
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 12,429 | 16.0 | 4. Face Detection | 5,464 | 64.9 |
| 2. Pose Estimation | 4,797 | 5.60 | 5. Object Detection | 1,737 | 137.8 |
| 3. Depth Estimation | 5,178 | 39.9 | 6. Image Segmentation | 2,705 | 43.8 |
OpenVINO NPU Benchmarks – Intel AI Boost
This OpenVINO-based benchmark evaluates the Intel AI Boost NPU on the ASUS ROG Zephyrus G16 for its AI processing capabilities, where it demonstrates high quantized performance, making it particularly effective for AI tasks benefiting from optimized quantized precision.
Important Note: To maximize NPU performance, it is recommended to install the latest drivers from Intel’s website. Detailed installation instructions are available in the Intel Release Notes.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 3,124
- Half Precision: 8,184
- Quantized: 11,822
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 14,834 | 19.1 | 4. Depth Estimation | 5,498 | 42.4 |
| 2. Pose Estimation | 5,267 | 6.15 | 5. Object Detection | 1,822 | 144.5 |
| 3. Face Detection | 5,696 | 67.7 | 6. Image Segmentation | 2,540 | 41.2 |
Intel Arc Graphics Benchmark (ONNX)
The Intel Arc Graphics integrated graphics benchmark provides basic AI processing capabilities suitable for lower-intensity tasks, with a notable focus on Style Transfer and Depth Estimation. However, the RTX 4090 outperforms it in higher-demand AI tasks.
View the full benchmark test here.
Benchmark Scores
- Single Precision: 2,614
- Half Precision: 1,056
- Quantized: 2,045
Top AI Tasks (Single Precision)
| AI Task | Score | IPS | AI Task | Score | IPS |
|---|---|---|---|---|---|
| 1. Style Transfer | 33,724 | 43.4 | 4. Depth Estimation | 5,954 | 45.9 |
| 2. Pose Estimation | 11,926 | 13.9 | 5. Image Segmentation | 1,353 | 21.9 |
| 3. Image Super-Resolution | 5,718 | 211.2 | 6. Face Detection | 1,263 | 15.0 |
Final Summary
The ASUS ROG Zephyrus G16 with Intel Core Ultra 9 185H, NVIDIA GeForce RTX 4090, and Intel AI Boost NPU offers outstanding AI performance, particularly in Style Transfer, Pose Estimation, and Depth Estimation. The RTX 4090 provides efficient, high-throughput GPU processing for complex tasks, while the Intel Core Ultra 9 185H CPU delivers balanced performance for moderate AI workloads. Additionally, the Intel AI Boost NPU enhances efficiency in quantized precision tasks, making the system highly adaptable for a range of AI applications. This setup is an excellent choice for users requiring both GPU and CPU-NPU AI processing capabilities in a mobile, high-performance form factor.