New Procyon AI Benchmark for Macs now available
June 25, 2025
We’re excited to announce you can now use Procyon to measure the AI image generation performance of Apple Macs using the CoreML AI Inference engine.
As easily creating high-quality generated images has become more accessible to everyone, many will have seen AI-generated images on social media to create memes, thumbnails, and edited photos.
However, AI image generation has many benefits to offer for enterprises. With only text prompts, processes like mockup creation, visualizing concepts or finding temporary images for testing content layouts have become considerably quicker.
The Procyon AI Image Generation Benchmark measures how quickly your devices complete these AI tasks locally – without the need for an online service. It supports AI Inference technologies from almost every major player in the industry, and we’re excited to add Apples’ CoreML to the list of supported inference engines today.
With CoreML, you are able to run the Image Generation Benchmark Stable Diffusion 1.5 and XL tests on Apple-Silicon-based macOS devices at FP16 precision. As with all other inference engines, the performed workload is the same, and results are comparable with results from other inference engines when using the same AI model and Precision.
Alongside this update, we are also enabling the ability for users to upload their Procyon Image Generation results using the ‘--online' flag – a first for one of our Procyon Benchmarks for macOS. This optional feature requires a 3DMark.com account, and just like with 3DMark and Procyon on Windows, you’re able to conveniently store your results and all generated images online for future reference. Uploaded results are hidden from the public by default, but there is an option to reveal them for sharing and comparing with other Image Generation scores.
You can see an example visible score here - https://www.3dmark.com/procyon/135892725
We have also added the option to export all generated images as a collection of .png images with the ‘--export-images’ flag, and the entire result as a .pdf with ‘--pdf-export’.
Available now
This Procyon AI Image Generation Benchmark for macOS is available now for any organization or individual with an active Procyon AI Inference Benchmarks license. To download the Benchmark, please get in contact with us.
If you’re interested in learning about how Procyon AI Inference benchmarks can assist your business, you can reach out to us at https://benchmarks.ul.com/contact.
Members of the press can get in contact with us at https://benchmarks.ul.com/press
Recent news
-
New Procyon AI Benchmark for Macs now available
June 25, 2025
-
3DMark for macOS available now!
June 12, 2025
-
New Inference Engines now available in Procyon
May 1, 2025
-
Try out NVIDIA DLSS 4 in 3DMark
January 30, 2025
-
Test LLM performance with the Procyon AI Text Generation Benchmark
December 9, 2024
-
New DirectStorage test available in 3DMark
December 4, 2024
-
New Opacity Micromap test now in 3DMark for Android
October 9, 2024
-
NPUs now supported by Procyon AI Image Generation
September 6, 2024
-
Test the latest version of Intel XeSS in 3DMark
September 3, 2024
-
Introducing the Procyon Battery Consumption Benchmark
June 6, 2024
-
3DMark Steel Nomad is out now!
May 21, 2024
-
Procyon AI Inference now available on macOS
April 8, 2024
-
Procyon AI Image Generation Benchmark Now Available
March 25, 2024
-
Announcing the Procyon AI Image Generation Benchmark
March 21, 2024
-
3DMark Steel Nomad will be free for all 3DMark users.
December 20, 2023