The M5 Ultra Mac Studio tears through our benchmark tests
First reported by The Verge ·
The M5 Ultra Mac Studio significantly outperforms its predecessor, accelerating professional workflows for AI and visual effects.
Apple has released its new Mac Studio, featuring the M5 Max and M5 Ultra chips. The M5 Ultra model tested, equipped with a 36-core CPU, 80-core GPU, 256GB of RAM, and 4TB of storage, costs $12,299. This high-end configuration is aimed at demanding users like AI developers and visual effects houses, exceeding typical content creation needs. Shipping begins today, with the M5 Max model starting at $2,499 and the M5 Ultra at $5,499. Preliminary benchmarks show the M5 Ultra Mac Studio is approximately 30% faster than its M3 Ultra predecessor in CPU tasks and significantly faster in graphics, with a 63% uplift in Cinebench. Storage speeds are also exceptionally high, and tasks like 4K video exports are completed in under a minute.
The performance leap in the M5 Ultra Mac Studio, particularly in CPU and GPU-intensive tasks, signals Apple's continued commitment to pushing the boundaries for high-performance computing within its ecosystem. This advancement is crucial for professionals in fields like AI development and 3D rendering, who require immense processing power for complex simulations and data analysis. The substantial increase in speed means these users can potentially reduce project completion times and tackle more ambitious projects, directly impacting productivity and innovation.
This new generation of Mac Studio, especially with the M5 Ultra chip, is positioning itself as a key tool for the burgeoning AI sector. By offering such significant performance gains, Apple is making a strong case for its hardware to be the platform of choice for developing and deploying AI models locally. Users engaged in these cutting-edge fields should anticipate more seamless integration and faster iteration cycles, while competitors will face increased pressure to match these capabilities.
AI-written summary. May contain errors.