Frame Desktop Review: Surprise

In the 3Dmark Steel Nomad, a popular graphics benchmark, frame desktops execute 13% faster than the ROG Flow Z13. This is also the best performance mode for both systems. It still lags behind the latest generation of desktop GPUs, such as the RTX 5060, but that’s expected. The closest discrete GPU to be comparable is the laptop RTX 4060. While this isn’t exciting for PC gamers, remember: these are integrated graphics. The only close integrated graphics performance is the Apple side of the M4 chip.
I’ve tried some modern games like Cyberpunk 2077 and Marvel competitors, Their performance impressed. I was surprised to find that I could play Cyberpunk 2077 There is no upgrade at 1080p and Ultra settings, but it reaches 84 frames per second on average. If you use AMD’s image upgrade settings or even frame generation, you can easily venture into higher resolutions and even ray tracing. Lighter games Marvel competitors Play football without missing it.
Due to the nature of the chip, although it does have a large capacity option, it is not removable RAM. The base model has 32 GB, while my configuration has 64 GB. There is a 128 GB option, which is important for anyone who wants to work with a large local language model or other AI projects. The internal neural processing unit is able to have more than 50 tops, and the framework claims it can run large language AI models in real time, such as the Llama 3 70B native, especially on the 128-GB model, which can enable GPU 96 GB of VRAM to be used.
Niche Market
Photo: Luke Larsen
Frame laptops are already pretty good, but frame desktops are also probably the most niche the company has launched so far. The benefits of choosing a frame desktop, rather than other desktops, are related to what makes the frame spirit so unique. There are even smaller options, such as the Corsair AI workstation 300, which are slightly taller but not as deep. They are priced the same, although Corsair sells only 128 GB of RAM for a premium option.
Frame desktop, on the other hand, is a wonderful middle ground between the compact nature of a mini PC and the performance and upgradeability of a true small-part PC. I don’t like that you can’t upgrade RAM, I want the side panel to be easier to install. I’m also worried about the internal CPU. Although outstanding, it depends largely on AMD’s continued support for this somewhat professional component. The framework hasn’t disappointed me so far, but its previous computers used more standard parts. But by now, the company has gained my trust.
For AI enthusiasts who want to run native LLM, the framework desktop is a shoo-In, especially in all available memory. But it’s also a good choice for those who want a compact PC that doesn’t completely confiscate performance or upgradeability.