AMD Radeon 860M vs Intel Arc Graphics 140T

AMD Radeon 860M

The Radeon 860M is a fairly powerful RDNA 3.5 architecture iGPU that debuted in early 2025 in Krakan Point APUs and a direct successor to the Radeon 760M. It packs just a few differences under the hood, such as faster caches; its 8 CUs/WGPs (512 unified shaders) run at up to 3,000 MHz.

DX12 Ultimate, ray tracing, AI image generation and other modern features are all supported here, as are many popular video codecs including first and foremost AVC, HEVC, VP9 and AV1. The latest addition to that list, the VVC codec, is not HW-supported unlike it is with Intel Lunar Lake chips. Miracast and FreeSync support is present. The iGPU can drive up to 4 monitors simultaneously with resolutions as high as SUHD 4320p60.

RDNA 3.5-based GPUs are projected to be fully supported by AMD until some time in 2027.

Performance

Its gaming performance is good enough for playing 2024 games in low graphics settings.

Power consumption

Krakan Point chips are a single-die design, meaning the iGPU is built with the same TSMC N4P process as the CPU cores for above average, as of late 2024, power efficiency.

Intel Arc Graphics 140T

Intel Arc 140T is the iGPU used in the Intel Arrow Lake H/HX processor series. It is based on the Xe+ architecture (and not the newer Xe2 architecture that e.g. the Arc Graphics 140V in Lunar Lake uses). The graphics adapter features 8 cores (128 CUs, 8 Raytracing units) with XMX instructions for improved AI performance and 8 MB L2-Cache.

Thanks to the similar core count and clock speed but higher TDP, the performance should be similar to the Xe2 based Intel Arc 140V iGPU. First leaked benchmarks even showed a vastly better performance. All in all, we think that the Arc 140T will be only suited for entry level gaming in lowest settings.

The Arc 140T is integrated in the H and HX series of CPUs but use an own chiplet that is manufactured at TSMC in 5nm (N5P process).