Primate Labs Releases Geekbench AI 1.2 Update
The latest version of the AI benchmark software introduces framework upgrades for Windows and Android systems.
Updated on Oct. 15, 2026 in Artificial Intelligence
Live Poll
Do you trust benchmark scores to accurately reflect your device's AI performance?
Primate Labs has officially released Geekbench AI 1.2, an update designed to improve workload performance for machine learning tasks. The new version incorporates key upgrades to several frameworks and backends.
Why it matters
These upgrades were implemented to enhance the accuracy and performance of artificial intelligence benchmarks on both Windows and Android systems. The update ensures that modern hardware can be more precisely measured during complex machine learning workloads.
The version 1.2 release integrates ArmNN 26.01 and Samsung ENN 3.1.13 alongside updated ONNX model quantization. These technical refinements specifically improve workload scoring for Windows users.
The players
Primate Labs
Primate Labs is a software development company that specializes in benchmarking tools for testing processor and system performance.
The details
The update improves the software's ability to measure CPU, GPU, and NPU performance by incorporating upgrades to the ONNX Runtime, OpenVINO, Samsung ENN, and Qualcomm QNN backends. By updating model quantization methods, the benchmark provides more reliable evaluations across its suite of computer vision and language tests.
Timeline
October 15, 2026: Primate Labs announced the availability of Geekbench AI 1.2.
The Tech Race
Geekbench AI remains a standard tool in a market defined by the rapid hardware innovation of specialized neural processing units. This release marks a shift toward optimizing benchmark frameworks to keep pace with the complex quantization requirements of modern AI models.
Users can now leverage more accurate performance data to better understand how their devices handle AI-driven tasks like computer vision. Developers and hardware enthusiasts should note that scores from version 1.2 are not comparable to previous iterations.
The takeaway
Users looking to compare device performance should ensure they are using the latest version of the benchmarking software to capture accurate data. Maintaining updated testing tools is essential for understanding how hardware handles emerging AI applications.
Further reading
For more context on how current benchmarks are evolving, see the latest updates in Artificial Intelligence.
More information
Download the software or view technical specifications on the Geekbench AI download page.
Source note: This article includes information reported by TechSpot.
Live Poll
Do you trust benchmark scores to accurately reflect your device's AI performance?
