Geekbench 7 launches with Next-Gen AI and Media Benchmarks

Primate Labs has officially released Geekbench and it introduces cross-platform benchmarking suite for smartphones, laptops, and desktop computers. The new version features updated CPU and GPU workloads designed to test modern artificial intelligence, machine learning, and high-performance media processing tasks.

New AV1, Whisper AI, and Smarter Multi-Core Benchmarking

Geekbench 7 adds new CPU workloads including AV1 screen sharing, Opus audio compression, Whisper live caption decoding, and Jolt engine game physics. The multi-core CPU benchmark has been redesigned so that workloads only run multi-threaded if real-world applications utilize multiple CPU cores for those specific tasks. On the graphics front, the GPU benchmark suite now adds official CUDA API support alongside OpenCL, Vulkan, and Metal. It introduces machine learning workloads such as face tracking filters, super-resolution upscaling, background blur, RAW image processing, and ray-traced path tracing.