Austin, Texas – The recent leaks of the AMD Ryzen AI 9 HX 370 “Strix Point” APU benchmarks reveal significant improvements in both CPU and GPU performance with the implementation of Zen 5 and RDNA 3.5 cores. This APU is part of the Ryzen AI 300 series and boasts a 12-core, 24-thread configuration with four Zen 5 and eight Zen 5C cores. Running at boost clocks of up to 5.1 GHz, the APU offers 36MB of cache and the Radeon 890M iGPU with 16 compute units or 1024 cores. Compared to its predecessor, the Ryzen 9 8945HS, the Ryzen AI 9 HX 370 delivers 50% more cores/threads, 33.3% more compute units, and 3.12x the NPU performance, showcasing significant generation-over-generation improvements.
The early Geekbench leak showcasing the Ryzen AI 9 HX 370 APU indicates that it is still an early sample given that the boost clock recorded is around 4.2 GHz, below its top-rated 5.1 GHz clock rate. Despite this, the CPU performance of the Zen 5 cores on the APU shows promising results.
Moreover, the leaked benchmark results of the Ryzen AI 9 HX 370 APU in Geekbench 5.4.5 and 6.3.0 demonstrate impressive performance gains. In Geekbench 5.4.5, the APU scored 1847 points in single-core and 14,316 points in multi-core tests. On the other hand, in Geekbench 6.3.0, it achieved 2544 points in single-core and 14,158 points in multi-core tests. Additionally, the RDNA 3.5 iGPU scored 41,995 points in the OpenCL test, highlighting the APU’s graphics capabilities.
The Ryzen AI 9 HX 370 APU surpasses the flagship Ryzen 9 8945HS APU in both single-core and multi-core performance, showcasing a 7% lead in single-core, a 20% lead in multi-core, and a significant 40% gain in the graphics test, putting it on par with the RTX 2050 discrete GPU.
With the upcoming release of the Ryzen AI 300 series in July 2024 through various OEM partners, consumers can anticipate improved availability starting in the holiday 2024 season or Q4. AMD’s decision to eliminate set TDP for the Ryzen AI 300 series provides manufacturers with more flexibility in choosing the TDP target for their devices. This approach allows the chips to scale from 15W up to 54W, potentially leading to enhanced performance across different power profiles.