AI accelerator Neural Processing Unit NPU, Digital signal processor DSP trillion operations per second
Apple A10 2016 iPhone 7/iPad 6/7 A10X 2017 iPad Pro 2
Apple A11 2017 iPhone 8/8 Plus/X ca, Neural Engine, 600 GOPS
Apple A12 2018 iPhone XS/XR cw/iPad 8/iPad Air 3/A12X iPad Pro 3 ce/A12Z 2020 iPad Pro 4, Next-generation Neural Engine, 8 cores, 5 TOPS
Apple A13 2019 iPhone 11/iPad 9 ks, 8-core, 6 TOPS
Apple A14 2020 iPhone 12 mini jl/iPad 10 jl/iPad Air 4 cw, 4-6 GB, 16-core, 11 TOPS
Apple M1 2020 iPad Air 5/iPad Pro 5 ca, 8- GB, 16-core, 11 TOPS
Apple A15 2021 iPhone 13/13 Pro/14 kt, 4-6GB, 16-core, 15.8 TOPS
Apple M2 iPad Air M2 kt/iPad Pro 6, 8- GB, 16-core, 15.8 TOPS
Apple A16 2022 iPad A16/iPhone 14 Pro/15 ca, 6 GB, 16-core, 17 TOPS
Apple A17 Pro 2023 iPhone 15 Pro, 8 GB, 16-core, 35 TOPS
Apple M3 2023 iPad Air M3/Mac 8GB, 16-core, 18 TOPS
Apple A18 2024 iPhone 16 jl, 35 TOPS
Apple M4 2024 iPad Pro 7, 8GB, 38 TOPS
Apple A19 2025 iPhone 17, ?? TOPS
Qualcomm Snapdragon systems on chips: Snapdragon_8/8s Gen 3 (2024)
Google Tensor with Tensor Processing Unit (Edge_TPU) used in Pixel Phones
A guide to AI TOPS and NPU performance metrics
* Qualcomm Snapdragon X series Elite/Plus DSP 45 TOPS, GPU 3.8-4.6 TFLOPS
* AMD Ryzen AI 9 HX Series 300, Ryzen AI 300 series, 365/370/375 28W (15-54W) 50 TOPS, 375 55 TOPS 375 vs 258V
* Intel Core Ultra 2 200V Series Intel Core Ultra 17W (8-17-37W): 2x6V 16GB, 2x8V 32GB; Core Ultra 5 228V 238V 8MB cache 40 TOPS, Ultra 7 258V 268V 12 MB cache 47 TOPS, Ultra 9 288V 30W 48 TOPS
Intel AMX XMX Advanced Matrix Extensions in CPU, Xe Matrix Extensions in GPU
Microsoft Copilot+ PC 40 TOPS, 16 GB memory, 256 GB storage
Apple Intelligence ChatGPT-4o Code Completion 16GB
Best chatbot ChatGPT alternatives Anthropic Claude, Google Gemini for live data, Microsoft Copilot creative, Perplexity research, Inflection Pi personal, xAI Grok social, Llama open source, MetaAI fun, Poe more choices comparison
全球43大AI工具排行榜，Perplexity、Gamma⋯有哪些必收好用AI App？ Perplexity, Claude Anthropic, ChatGPT OpenAI
