Download AI Benchmark APKs for Android

AdvertisementRemove ads, dark theme, and more with

Tools

What's new in AI Benchmark 7.0.0

1. New models: 1.1B Llama2 Large Language Model (LLM), Stable Diffusion V1.5 Model, SwinIR, Depth-Anything-V2, ConvNeXT-V2, MobileNet-V4, etc.
2. Added LiteRT V2 runtime (version 2.0.3).
3. LiteRT (TFLite) runtime updated to version 1.4.2.
4. Added an optional test checking the performance of 4-bit quantized LLMs with 16-bit activations.
5. Updated Qualcomm QNN, MediaTek Neuron, Samsung ENN, TFLite NNAPI and GPU delegates.
6. Updated in-app ranking table.

About AI Benchmark 7.0.0

Billion-Parameter LLMs, Neural Image Generation, Face Recognition, Image Classification, Question Answering...

AdvertisementRemove ads, dark theme, and more with

Premium

Is your smartphone capable of running the latest Deep Neural Networks to perform these and many other AI-based tasks? Does it have a dedicated AI Chip? Is it fast enough? Run AI Benchmark to professionally evaluate its AI Performance!

Current phone ranking: ai-benchmark.com/ranking

AI Benchmark measures the speed, accuracy, power consumption and memory requirements for several key AI, Computer Vision and NLP models. Among the tested solutions are LLMs, Neural Image Generation, Image Classification and Face Recognition methods, neural networks used for Image / Video Super-Resolution and Photo Enhancement, as well as AI solutions used in autonomous driving systems and smartphones for real-time Depth Estimation and Semantic Image Segmentation. The visualization of the algorithms’ outputs allows to assess their results graphically and to get to know the current state-of-the-art in various AI fields.

In total, AI Benchmark consists of 83 tests and 30 sections listed below:

Section 1. Classification, MobileNet-V4
Section 2. Classification, EfficientNet-B4
Section 3. Face Recognition, Swin Transformer
Section 4. Classification, ConvNeXT-V2
Section 5. Classification, MobileViT-V2
Section 6. Optical Character Recognition, ViT Transformer
Sections 7/8. Parallel Model Execution, 8 x EfficientNet-B4
Section 9. Object Tracking, YOLO-V8
Section 10. Semantic Segmentation, DeepLabV3+
Section 11. Parallel Segmentation, 2 x DeepLabV3+
Section 12. Semantic Segmentation, Segment Anything
Section 13. Photo Deblurring, IMDN
Section 14. Image Super-Resolution, SwinIR
Section 15. Image Super-Resolution, SRGAN
Section 16. Image Denoising, SplitterNet
Section 17. Depth Estimation, Depth-Anything-V2
Section 18/19. Image Enhancement, DPED
Section 20. Learned Camera ISP, PyNET
Section 21. Bokeh Effect Rendering, U-Net
Section 22/23. 4K Video Super-Resolution, XLSR
Section 24/25. 4K Video Super-Resolution, VideoSR
Section 26. Neural Text Generation, GPT2
Section 27. Neural Text Generation, Llama2 1.1B
Section 28. Neural Image Generation, Stable Diffusion V1.5
Section 29. Neural Text Generation, Llama2 110M 4W16A
Section 30. Memory Limits, ResNet

Besides that, one can load and test their own TFLite deep learning models in the PRO Mode.

A detailed description of the tests can be found here: ai-benchmark.com/tests.html

Note: Hardware acceleration is supported on all mobile SoCs with dedicated NPUs and AI accelerators, including Qualcomm Snapdragon, MediaTek Dimensity / Helio, Google Tensor, HiSilicon Kirin, Samsung Exynos, and UNISOC Tiger chipsets. Starting from AI Benchmark v4, one can also enable GPU-based AI acceleration on older devices in the settings ("Accelerate" -> "Enable GPU Acceleration" / "Arm NN", OpenGL ES-3.0+ is required).

Show more Show less