AMD
GPUIntegrated (APU)Released Oct 15, 2024

AMD Xbox Series X GPU (6 nm refresh)

16 GB GDDR6 · 320-bit · 3,328 cores

This is the smaller, 6 nm version of the Xbox Series X processor, used in the console revisions Microsoft released on 15 October 2024: the 2 TB Galaxy Black Special Edition and the white all-digital model without an optical drive. The GPU configuration is identical to the launch console at 52 enabled RDNA 2 compute units and a fixed 1.825 GHz, so performance and compatibility are unchanged; the redesigned board with the smaller chip draws slightly less power.

General

Manufacturer
AMD
Model
Xbox Series X GPU (6 nm refresh)
Segment
Integrated (APU)
GPU chip
Scarlett (6 nm)
Architecture
RDNA 2
Release date
Oct 15, 2024

Memory

Memory size
16 GB
Memory type
GDDR6
Bus width
320-bit
Bandwidth
560 GB/s
Memory clock
1,750 MHz

Clocks & silicon

GPU clock
1,825 MHz
Boost clock
—
Shader units
3,328 Stream Processors (52 CU)
Tensor / AI cores
—
RT cores
52
Transistors
—
Technology (nm)
6 nm
TSMC N6
TDP / TBP
—

Interface & APIs

Bus interface
Integrated (on-die)
SLI / multi-GPU
—
DirectX
12 Ultimate (12_2)
OpenGL
—
Vulkan
—
Other APIs
Xbox Direct3D 12.X

AI support

Precisions
FP8FP16BF16INT8INT4TF32FP64
AI accelerator
None (RDNA 2 packed math / DP4a integer dot product)
Peak AI TOPS
49 TOPS
Notes
RDNA 2 has no dedicated matrix or AI accelerator. Microsoft rates the GPU at 49 TOPS for INT8 and 97 TOPS for INT4 using the compute units' dot-product instructions, and packed FP16 runs at twice the FP32 rate. There is no BF16, TF32 or FP8 support.
All AMD GPUs →

Integrated (APU) gpus from AMD announced around the same time as the AMD Xbox Series X GPU (6 nm refresh). Add any of them to your compare list to see the specs side by side.