Skip to content
Search

Latest Stories

Apple M5 with Neural accelerators and 16-core Neural engine powers next-level AI

Apple’s M5 chip represents the next big step in performance for Apple silicon

Apple M5

M5 chip delivers over 4x peak GPU compute performance for AI

Apple

Highlights

  • M5 chip delivers over 4x peak GPU compute performance for AI compared to M4 and doubles AI performance over M1.
  • Features next-generation 10-core GPU with Neural Accelerator in each core, faster CPU, 16-core Neural Engine, and 153GB/s unified memory bandwidth.
  • Powers new 14-inch MacBook Pro, iPad Pro, and Apple Vision Pro with improved AI workflows, graphics, and energy efficiency.

Next-generation GPU and AI performance

Apple’s M5 chip represents the next big step in performance for Apple silicon, built on third-generation 3-nanometre technology. The chip introduces a 10-core GPU architecture, each core featuring a dedicated Neural Accelerator, enabling AI workloads to run dramatically faster.

According to Apple, the M5 delivers over 4x peak GPU compute performance compared to M4 and over 6x for AI performance compared to M1. The GPU also features third-generation ray tracing, delivering up to 45 percent higher graphics performance, while enhanced shader cores and dynamic caching provide smoother rendering and more realistic visuals for creative and gaming applications.


Applications like Draw Things and AI platforms such as webAI benefit from faster processing of diffusion models and local large language models on the 14-inch MacBook Pro and iPad Pro. Apple Vision Pro can render more pixels at up to 120Hz refresh rates, reducing motion blur and enhancing visual fidelity.

Apple M5 The 16-core Neural Engine accelerates on-device AIApple

Faster CPU, Neural Engine, and memory

M5 also features up to a 10-core CPU with six efficiency cores and four performance cores, delivering 15 percent faster multithreaded performance than M4. The 16-core Neural Engine accelerates on-device AI, powering tasks like transforming 2D photos into spatial scenes or generating Personas on Apple Vision Pro with greater speed and energy efficiency.

The chip offers a unified memory bandwidth of 153GB/s, a nearly 30 percent increase over M4, allowing larger AI models to run entirely on-device. With up to 32GB memory capacity, M5 supports demanding creative workflows such as simultaneous use of Adobe Photoshop and Final Cut Pro, alongside background cloud uploads.

Optimised for creative workflows and energy efficiency

Apple’s M5 is engineered to integrate seamlessly with Apple software frameworks, including Core ML, Metal 4, and Metal Performance Shaders, allowing developers to leverage Neural Accelerators directly. Creative applications, games, and AI-powered tools experience immediate performance boosts without additional programming.

The M5’s efficiency contributes to Apple’s 2030 carbon-neutral plan, reducing energy consumption across the chip’s lifetime while delivering world-class performance on the 14-inch MacBook Pro, iPad Pro, and Apple Vision Pro.

Pre-orders for devices powered by M5 are available now, offering users faster AI capabilities, graphics performance, and multitasking power in a single chip.

More For You

Google's Gemini

Apple is reportedly close to finalising a deal with Google

Getty Images

Apple calls on Google's Gemini to rescue Siri with 1.2 trillion-parameter AI deal

Highlights

  • Apple set to license Google’s Gemini AI model, which boasts 1.2 trillion parameters
  • Deal reportedly worth around $1 billion per year
  • Gemini will handle complex Siri functions until Apple’s own AI is ready

Apple strikes temporary deal with Google

Apple is reportedly close to finalising a deal with Google to power Siri using the tech giant’s advanced AI model, Gemini. The agreement, valued at around $1 billion per year, marks one of the biggest collaborations between two of Silicon Valley’s fiercest competitors.

The Gemini model, which has approximately 1.2 trillion parameters, will be integrated into Siri to handle tasks such as summarising information, multi-step planning, and contextual understanding , areas where Apple’s voice assistant has historically fallen behind its rivals.

Keep ReadingShow less