AMD Ryzen AI Max+ 395

The AMD Ryzen AI Max+ 395 is a 16-core, 32-thread mobile processor built on the Zen 5 architecture (Strix Halo) at a 4 nm process node, with a 55 W TDP, 64 MB of shared L3 cache, and a Radeon 8060S integrated GPU. It sits in the 96th percentile of all CPUs in the benchmark database, with an average benchmark score of 75,265, placing it among the top 4% of processors tested. Its nearest rivals—Intel Core Ultra 9 285, AMD EPYC 8224P, Intel Core Ultra 9 275HX, and AMD Ryzen 9 8945HX—all fall within a narrow performance band, with deltas ranging from -1% to +1.6%, indicating that the Ryzen AI Max+ 395 is highly competitive in its class.

Benchmark Performance

The Ryzen AI Max+ 395 delivers strong multi-threaded throughput. In Cinebench R23, it scores 45,228 points in the multi-core test and 6,385 points in the single-core test. The multi-core score is the headline figure, but the single-core result is also robust, reflecting a well-balanced design. The PassMark suite reinforces this: the processor achieves a multithread score of 53,137 and a single-thread score of 4,123. These numbers, when compared to its nearest rivals, show that the Ryzen AI Max+ 395 is essentially tied with the Intel Core Ultra 9 285 (average score 75,400, delta -0.2%) and the AMD EPYC 8224P (75,582, delta -0.4%), while trailing the Intel Core Ultra 9 275HX by 1% (76,024) and leading the AMD Ryzen 9 8945HX by 1.6% (74,103). The differences are small—less than two percentage points across the entire rival set—meaning that in real-world applications, the Ryzen AI Max+ 395 will be indistinguishable from these competitors in most workloads. The PassMark sub-scores further illustrate its capabilities: integer math at 195,223, floating-point math at 123,833, data compression at 656,891, and data encryption at 34,086. These are all high absolute values, and the extended instructions score of 52,958 indicates solid SIMD and AVX performance. The physics score of 3,448 and the random string sorting score of 72,894 round out a well-rounded profile. The average benchmark score of 75,265 is within 0.2% of the Intel Core Ultra 9 285 and 0.4% of the EPYC 8224P, confirming that the Ryzen AI Max+ 395 is a top-tier mobile processor.

Power and Thermals

The 55 W TDP is a defining characteristic of this processor. For a 16-core, 32-thread part built on a 4 nm process, this is a remarkably low power envelope. The die size is listed as 2x 70.6 mm², which, combined with the TSMC 4 nm node, suggests high power efficiency. The low TDP means that cooling requirements are modest; a standard laptop or compact desktop cooling solution designed for 55 W-class processors should suffice. There is no need for exotic liquid cooling or oversized heatsinks. This is particularly relevant for mobile form factors, where thermal constraints are tight. The fact that the processor maintains competitive performance at this TDP indicates excellent architectural efficiency. The memory subsystem—LPDDR5X at quad-channel with 256.0 GB/s bandwidth—also contributes to overall system efficiency, as high-bandwidth memory reduces the need for excessive cache or clock speeds. The 64 MB shared L3 cache further helps by reducing memory traffic. In summary, the Ryzen AI Max+ 395 is a low-power, high-efficiency part that should run cool and quiet in typical usage, making it suitable for thin-and-light laptops and other thermally constrained devices.

Single-Thread vs Multi-Thread Behavior

The split between single-thread and multi-thread performance reveals a processor that excels in both domains, though its multi-thread scaling is not linear. In Cinebench R23, the multi-core score of 45,228 is roughly 7.1 times the single-core score of 6,385. Given 16 cores and 32 threads, perfect scaling would yield a 32x improvement, but real-world scaling is limited by memory bandwidth, cache coherence, and software parallelism. A 7.1x ratio is typical for a 16-core/32-thread processor, indicating that the cores are well utilized but not saturating. The PassMark results show a wider gap: multithread 53,137 versus single-thread 4,123, a ratio of about 12.9x. This is because PassMark's multi-thread test often scales better with thread count than Cinebench, which has more serial sections. The single-thread scores themselves are strong—6,385 in R23 and 4,123 in PassMark—placing this chip in the upper echelon for single-threaded tasks like web browsing, office productivity, and light gaming. The data suggests that the Ryzen AI Max+ 395 will handle both lightly threaded and heavily threaded workloads with ease. For example, the PassMark integer math score of 195,223 and floating-point math of 123,833 indicate strong sustained throughput in parallel compute tasks, while the single-thread score of 4,123 ensures snappy responsiveness in everyday applications. The balanced nature of this processor means it does not sacrifice one type of performance for another, making it a versatile choice for users who run a mix of workloads.

How It Compares

Intel Core Ultra 9 285 — The Ryzen AI Max+ 395 trails the Intel Core Ultra 9 285 by a negligible 0.2% in average benchmark score (75,265 vs. 75,400). This is effectively a tie. Both processors deliver nearly identical multi-threaded and single-threaded performance, and the difference is within run-to-run variance. In practical terms, users will not notice any difference between the two in real applications.

AMD EPYC 8224P — The EPYC 8224P, a server-oriented part, edges out the Ryzen AI Max+ 395 by 0.4% (75,582 vs. 75,265). Despite being a different market segment, the EPYC's higher core count (though not listed in the fact pack) and server optimizations give it a slight edge. However, the Ryzen AI Max+ 395 is a mobile processor with a 55 W TDP, making this a remarkable result—it nearly matches a server chip while consuming far less power.

Intel Core Ultra 9 275HX — The Ryzen AI Max+ 395 is 1% behind the Intel Core Ultra 9 275HX (76,024 vs. 75,265). This is the largest deficit among the rivals. The 275HX is a high-end mobile part, and the 1% gap is still very small, but it indicates that the Intel chip has a slight performance advantage in aggregate. The Ryzen AI Max+ 395 compensates with its lower TDP and integrated GPU, which may be more important for certain use cases.

AMD Ryzen 9 8945HX — The Ryzen AI Max+ 395 outperforms the Ryzen 9 8945HX by 1.6% (75,265 vs. 74,103). This is the only rival it leads, and the margin is the largest among the four comparisons. The 8945HX is a previous-generation high-end mobile processor, so the newer Strix Halo design shows a clear generational improvement. The Ryzen AI Max+ 395 achieves this lead while likely consuming less power (though the 8945HX's TDP is not listed, the 55 W TDP of the Max+ 395 is notably low for a 16-core part).

Platform and Compatibility

The Ryzen AI Max+ 395 uses AMD Socket FP11, a mobile-specific socket. It supports LPDDR5X memory in a quad-channel configuration, delivering a memory bandwidth of 256.0 GB/s. This high bandwidth is crucial for both the CPU cores and the integrated Radeon 8060S GPU, which shares system memory. ECC memory is supported, which is unusual for a mobile processor and adds reliability for workstation-like tasks. The PCIe interface is Gen 4 with 16 lanes (CPU only), providing ample bandwidth for discrete GPUs or high-speed NVMe storage, though the integrated GPU may reduce the need for a discrete solution. The processor is based on the Zen 5 architecture, codenamed Strix Halo, and is manufactured on a 4 nm TSMC process. The die size is 2x 70.6 mm², indicating a chiplet design, though the exact layout is not specified. The production status is listed as Active, and the part number is 100-000001099. The multiplier is not unlocked, so overclocking is not supported. The memory bus being quad-channel is a standout feature, as most mobile processors use dual-channel; this contributes to the high bandwidth and likely helps the integrated GPU perform closer to a discrete entry-level card. The upgrade path is inherently limited due to the socket being mobile-specific; users cannot swap processors in most laptops. However, the integrated Radeon 8060S GPU means that the platform can handle graphics tasks without a separate card, making it a complete package for thin-and-light devices. The PCIe Gen 4 support ensures compatibility with current peripherals, and the 16 lanes allow for a fast discrete GPU if needed. Overall, the platform is modern and well-equipped for a high-end mobile system, with the caveat that the socket is not upgradeable.