Geekbench 7 is out with overhauled performance tests for CPU and GPU workloads

Primate Labs released **Geekbench 7**, a major update to its widely used cross-platform benchmark suite. The release overhauls performance tests for both **CPU** and **GPU** workloads. Coverage from **9to5Mac** frames the launch around a clear goal: scores that track how people actually use phones, laptops, and desktops rather than synthetic loads that drift from real software.

The technical pitch is an overhaul of the test design itself. **Geekbench 7** targets CPU and GPU paths that map to everyday tasks, gaming, and AI workloads. That mix matters because modern devices share silicon across UI work, media, games, and on-device models. A single-number suite that only stress-tests raw compute can mis-rank chips that win on one path and lose on another.

For engineers and builders, a closer match to daily use changes how you read results. When you pick silicon for a product, compare SKUs, or set performance budgets, you care about the mix of work your users hit—not only peak throughput on an artificial loop. An overhaul aimed at everyday tasks, gaming, and AI is useful when those three dominate the device’s real duty cycle.

In the market, **Geekbench** remains a common language for comparing Apple, Android, Windows, and other platforms in reviews and launch coverage. **Primate Labs** is not inventing the category; it is refreshing the yardstick competitors and OEMs already cite. When the tests shift, historical score charts break, and vendors that optimized for the old suite may look different under the new one—especially where GPU and AI paths were thin or missing before.

Watch how early **Geekbench 7** runs land across flagship phones and laptops, and whether review sites and OEMs adopt the new scores as the default reference. Treat first-wave numbers as a new baseline, not a direct drop-in for **Geekbench** history, until the community settles on which sub-scores matter for everyday, game, and AI-heavy use.