Framework Desktop Review: A Delightful Surprise

3 Demark Steel Traveler, a popular graphics benchmark has performed the Framework Desktop Ortiji Floj 13 percent faster than Flow Z 13. It also sets the best performance mode on both systems. It is still 32 percent behind the latest generation of desktop GPUs, such as RTX 5060However, it is expected. The nearest comparable GPU is the laptop RTX 4060. Although it is not exciting for PC gamers, remember: these are integrated graphics. Turn around is the only integrated graphics performance turn around Apple sides including M4 ChipsThe

I’ve tried some modern games, as Cyberpunk 2077 And Marvel rivals, And it was completely impressed with how well they performed. I was surprised to find out that I could play Cyberpunk 2077 1080p and Ultra Settings still achieve 84 frames on average per second without any upscaleing. If you use AMD image upscaleing settings or even frame generation, you can easily enter higher resolution and even ray tracing. Like lighter games Marvel Play not getting a beat.

The RAM is not removable because of the nature of the chip, though it comes to large capacity options. The base model has 32GB, when my configuration has come up with 64GB. There is a 128GB option, which is important for anyone to work with local large language models or other AI projects. Enabled to over -processing units inside 50 topAnd framework claims that it can run large language AI models like Lama 3 70B locally in real time, especially in the 128-GB model that allows GPU 96GB VRAM to work.

A niche in a niche

Photograph: Luke Larson

Framework laptops were already great, but the Framework Desktop is probably the company’s most niche product launched so far. Framework does not have to do less with the benefits of choosing the desktop on top of any other desktop. There are more smaller options as Corsier AI Workstation 300Which is a bit taller but not as deep. They are similarly priced, though the cursor is selling only the high-end option with 128 GB RAM.

The framework desktop, on the other hand, is a great medium of the compact nature of a mini-PC and the actual small-factor PC performance and upgradible land. I don’t like that you can’t upgrade the RAM and I hope that the side panels were easy to install. I also have concerns about the CPU inside. Although it performs great, AMD depends a lot on this some specialized element. The framework did not let me down so far, but its previous computers used more standard parts. However, by now the company has gained some confidence from me.

For AI enthusiasts hoping to run local LLMs, the framework desktop is a shoe-in, especially with all the available memories. However, it is also a decent option for those who want a compact PC that does not completely seize performance or upgradibility.