M4 vs M5 MacBook Pro: Here’s how I made up my mind

Performance is the flagship feature for creative professionals. The faster you can render an animation file, export Adobe Lightroom photos, or do something crazy like create your own big language model for artificial intelligence, the more time you save. For most creatives, a faster, more efficient workflow can literally generate more income and reduce stress.

As a writer and photographer, I use a 14-inch MacBook Pro M3 daily and now is a good time to consider an upgrade. I know in the back of my mind that Apple has been pushing the boundaries lately, and I also know that the new M5 MacBook Pro is significantly faster.

How much faster? Apple claims that for AI tasks like the large language model database I mentioned, you can expect speed improvements on the order of 3.5x faster than the M4 version. The M5 is also 6x faster than the original M1 model, released in 2020. The new M5 14-inch MacBook Pro costs $1,599/£1,599, which isn’t exactly an impulse buy.

That’s why I decided to do some of my own workflow testing using apps like Blender and Adobe Lightroom. Next, I “crunched the numbers” using benchmarking apps like Geekbench AI and Blackmagic Disk Speed ​​Test. While I haven’t always been blown away by the performance improvements, there have been some shocking revelations. I used macOS 26.1 on both machines, and they both came with 16GB of RAM (coincidentally, both are Space Black).

I give my thoughts after explaining the results of each test, as well as whether I’m ready to jump into the new machine.

Creative workflows

To test the creative process, I chose tasks that tend to take time and slow me down. On my M3, I may have cursed Apple’s engineers a bit while waiting for my photos to export from Adobe Lightroom, for example. It seems like an eternity when an event planner texts and emails you repeatedly asking for final photos. I also created a song in GarageBand and mastered the audio file, then tried a Blender animation render.

I use Adobe Lightroom daily. This is my primary photo editing and management app, so I use it extensively and am very familiar with its features. I own a trusty Canon 5D Mark II, but was reviewing a higher end Nikon Z5 camera. I recently covered an event and took 562 photos with the Z5. After loading the photos onto both MacBooks, I imported the new photos, did some editing, then exported the results as JPG files to the local drive.

Overall, Lightroom looked snappy on the M5 compared to the M4. I remember making my edits and seeing the results appear instantly; with the M4 (and certainly with my own M3) there was often a slight delay at times. For mass export, the M5 MacBook Pro was about 10 seconds faster than the M4 model. The M5 recorded a time of 55.96 seconds to export all 562 photos; the M4 took 1 minute 6.4 seconds to export the same files.

I am also a musician and have been playing guitar for many years. I love that Apple still includes GarageBand on a new MacBook Pro like the M5 and I tend to get lost in the creative process. I added several drum tracks, piano and a few other instruments to a new song. To test the MacBooks, I duplicated the tracks to create a large file.

The test results were similar to what I experienced with Lightroom. My song exported about 10 seconds faster with the M5, which is a nice time saver if you create multiple audio files throughout the day. The M5 took 56.01 seconds to generate an AAC file, while the M4 took 1 minute 5 seconds. These real-world tests are more valuable to me than a benchmark application, which I’ll get to in a moment, because the numbers are more meaningful to me. I want to be able to do real work and then know how much time I would save with the new machine.

I can’t say that I use Blender regularly, but I have great aspirations to learn this application in more depth, especially to combat all the abuses of today’s AI. My goal is to learn how to make real animation work before it becomes a lost art form. This type of stress test is also suitable for MacBook Pro because it requires multiple files to be rendered. I used a Blender demo project that involves rendering dozens of 3D models.

Test results weren’t that different between the M4 and M5 laptops. The M5 returned the models in 9 hours 45 minutes while the M4 took 9 hours 55 minutes. It seemed like forever, but then I looked at how long rendering 3D models could take and sometimes it’s closer to an hour or more. However, I wasn’t as impressed with the speed difference between the M5 and M4 and began to wonder how much faster the new machine was.

John Brandon

Benchmarking Applications

In many ways, workflow testing is more subjective. The two MacBooks seemed similar in performance when I ran my own tests with Lightroom and GarageBand, and I wasn’t blown away by the speed differences. When I ran three benchmarking apps, I realized that there was a remarkable speed difference after all, especially during AI and GPU-related tasks. This is consistent with what Apple has said, that speed gains are linked to these activities. Note: For these benchmark tests, I will only report actual benchmarking results.

Blackmagic Disk Speed ​​Test

This application is intended to show you the performance of a local disk, but of course it also depends on the processor, which must process all file transfers. Blackmagic Disk Speed ​​Test is a free utility that indicates that the M5’s read and write speeds are significantly faster than those of the M4. My test results (which match those from the MacBook Pro M5 review) for the M5 for write speed was 6,440 MB/s, and for read speed it was 6,725.6 MB/s. For the M4, the write speed was only 3,265.3 MB/s and the read speed was 2,904 MB/s. This means that the M5’s write speed is 97% faster than M4 and read speed is 131% faster. That’s a big win.

Here is the most breathtaking result of all my tests. Geekbench lets you test AI processing speed for CPU, GPU, and Neural Engine. I chose GPU processing for this, but all results were better on the M5 than the M4. The quantified score is the average of single-precision and half-precision tests. The M5 quantified score was 23628, while the M4 quantified score was only 11616. This roughly represents twice as fast like the previous MacBook.

3DMark Extreme Solar Array

3DMark offers a free benchmarking app that lets you run a gaming test called Solar Bay Extreme. Basically, this is a GPU test that shows you frames per second (fps). The M5 recorded a rate of 101 fps on frame 300 and generally held around 100 fps for the duration of the test. The M4 ran at 70fps on frame 300 and through most of the test, dipping into the low 60s at times. This means the M5’s frame rate was about 50% faster than the M4 when rendering games. Gamers, you will love using a MacBook Pro M5.

John Brandon

My decision: MacBook Pro M5

I’m sold. During testing, I knew that benchmarking applications would reveal a major performance increase over the previous model. I wasn’t as impressed with the results in my own workflow, although saving 10 seconds for exports and renders isn’t exactly impressive. In the end, the M5 was at least 50% faster in some tests and even up to 2x faster than the M4. I could see myself taking the leap to see these performance gains.

You might find a really good deal on the M4 MacBook Pro, in which case this might just influence your decision: