Processors with higher AI performance (TOPS)

Discover processors with the highest artificial intelligence performance, measured in TOPS (Tera Operations Per Second). This ranking evaluates the chips' ability to efficiently execute AI models, ideal for tasks like image recognition and natural language processing. A high TOPS value indicates greater AI processing power, enabling superior performance in machine learning applications. Explore the most powerful options for computers, mobile devices, and other hardware.

204100% verified
  1. 1

    Intel Core Ultra 200V Series (Lunar Lake)

    175 Global Votes
    • Provides 100+ TOPS of AI performance

      (+4)

    The Intel Core Ultra 200V Series (Lunar Lake) delivers exceptional AI performance, featuring an NPU that achieves up to 48 TOPS and a combined platform capability exceeding 100 TOPS. This architecture provides significant acceleration for AI computations with impressively low power consumption, tripling the TOPS of the previous generation. Its optimized design for ultrathin systems ensures a smooth and efficient user experience in artificial intelligence applications.

  2. 2

    AMD EPYC 9755

    14 Global Votes
    • Delivers 1.45x performance of Xeon 6980P

      (+4)

    The AMD EPYC 9004 and 9005 processors are specifically engineered to accelerate AI workloads, delivering leadership performance and energy efficiency. They incorporate AVX-512 extensions that optimize deep learning and data processing tasks, resulting in faster execution of AI algorithms.

  3. 3

    Intel Core Ultra 7 270K Plus

    9 Global Votes
    • Features Intel AI Boost with 13 TOPS (INT8) peak

      (+4)

    The Intel Core Ultra 7 270K Plus is a high-performance desktop processor that offers substantial improvements in overall and multi-threaded performance. Its advanced architecture and increased E-cores contribute to greater efficiency in demanding workloads, including those that benefit from AI capabilities. While the context does not directly specify its AI TOPS, its positioning as a next-generation processor with a focus on robust performance suggests enhanced capability for handling artificial intelligence workloads.

  4. 4

    AMD Ryzen 9 9950X

    6 Global Votes
    • Fastest desktop processor tested

      (+4)

    The AMD Ryzen 9 9950X delivers exceptional performance in AI workloads, such as Stable Diffusion image generation and multi-modal AI pipelines. Its Zen 5 architecture and 16 cores/32 threads enable it to handle large intermediate model states and significantly accelerate local inference and training tasks.

  5. 5

    Qualcomm Snapdragon X2 Elite Processors

    0 Global Votes
    • Features world's fastest NPU for laptops

      (+4)

    The Qualcomm Snapdragon X Elite features the world's fastest NPU for laptops, capable of delivering up to 45 TOPS (trillions of operations per second) of AI performance. This processor is designed to run generative AI models with over 13 billion parameters on-device at blazing-fast speeds, significantly enhancing efficiency and accuracy in artificial intelligence tasks.

  6. All the rankings you can imagine

    Thousands of verified votes to discover the best. Your vote here counts

  7. 6

    AMD Ryzen AI 9 HX 370

    0 Global Votes
    • Features improved Radeon 890M graphics

      (+4)

    This processor stands out for its impressive AI processing capability, delivering up to 80 total TOPS, with 50 TOPS dedicated to its NPU. Its advanced architecture and the integration of the Radeon 890M make it ideal for AI-intensive workloads. Furthermore, it provides robust multi-threaded performance and notable power efficiency for laptops.

  8. 7

    Intel Xeon 6

    0 Global Votes
    • Achieved 1.9x geomean performance improvement in AI

      (+4)

    The Intel Xeon 6 integrates AI acceleration into every core, enabling it to power inference, training, and small generative AI workloads. It delivers double the performance for AI, HPC, and data analytics due to its increased core count and double the memory bandwidth.