Three Mark Steel Gypsy, in a famous graphics benchmark, the framework desktop performed 13 % faster than the ROG Flow Z13. It is with both systems that are also based on excellent performance mode. It is still 32 % behind the latest desktop GPU generation, such as RTX 5060, but it can be expected. The nearest comparison is the GPU laptop RTX 4060. While this is not interesting for PC gamers, remember: These are integrated graphics. Only integrated graphics performance is approaching Apple with M4 chips.
I tried some modern sports, like Cyberpank 2077 And Miracle rivals, And it was fully affected how much they performed. I was surprised to know that I could play Cyberpank 2077 With no advanced 1080p and Ultra settings still get an average of 84 frames per second. If you use AMD’s image promotion settings or even frame generation, you can easily operate high resolutions and even traces of ray. Liked Games like Miracle competitors Play without defeat.
Ram is not able to withdraw due to the nature of the chip, though it comes in great potential options. The base model has 32 GB, while I came up with 64 GB. Here is a 128GB option, which is important to everyone who wants to work with local language local models or other AI projects. The inside nerve processing unit is worth more than 50 tips, and the framework claims that it can operate in real time, such as Lama 3 70B, such as Lama 3 70B, especially on the 128-GB model that works to work with the GPU 96 GB VRAM.
A niche of a niche
Photo: Luke Larson
The framework laptops were already good, but the framework desktop is probably the most niche product of the company so far. The benefits of choosing the framework desktop on any other desktop makes the framework of morals just as unique. There are even smaller options, such as the Course AIwork Station 300, which is slightly taller but not so deep. They have a similar price, though the courser is selling only the high -end option with 128 GB Ram.
On the other hand, the framework desktop is a tremendous medium ground between the compact nature of a mini -PC and a real small form factor PC performance and upgrade. I don’t like you can’t upgrade RAM, and I wish it was easy to install side panels. I also have concerns about the CPU inside. Although it performs very well, it depends on that AMD supports this special ingredient to some extent. The framework has not allowed me to be disappointed yet, but its previous computers have used more standard parts. But so far, the company has gained some confidence from me.
For AI -eager people in the hope of running a local LLM, the framework desktop is a show, especially with all available memory. But it is also a decent option for people who want a compact PC that does not completely waste performance or upgraded.
