AMD Ryzen

AMD's Ian Weinberg reveals how a computer the size of a book can run a 300-billion-parameter AI model locally: AMD's Ian Weinberg reveals how the new Ryzen AI Halo can run a 300-billion-parameter AI model locally on a computer the size of a book: "Inside of it is our next generation Ryzen AI Max+ Pro 495 processor." "The headline feature of the 495 is 192 gigs of shared RAM between the CPU and the GPU. When you look at developers today doing local AI, what they run out of very often is memory. It’s not the compute side, it’s the memory side." "Of that 192, you can reserve 160 of it for dedicated GPU VRAM, and that unlocks 300 billion parameter models that can be run local. Like GLM 5.3 Flash can be run directly on this box." "A 300 billion parameter model runs on a box that is basically the size of a book. Less than two inches tall." @AMD

AMD's Ian Weinberg gives a live demo of the Ryzen AI Halo running multiple massive AI models simultaneously on a computer the size of a book: "This is the Qwen 3.8 Flash Next, about 120 billion traditional parameters plus another 56 or so of n-grams. It made a song for me, and I don’t love it. So, make it more like dancehall Robyn with a slower beat." @AMD