AMD Ryzen AI Max+ 392 vs Intel Core Ultra 5 245HX Comparison

Head-to-Head Benchmarks

The recorded data shows a clear split between these two mobile processors, with the AMD Ryzen AI Max+ 392 winning 6 of the 11 shared benchmark comparisons, while the Intel Core Ultra 5 245HX takes 5 wins. The margin of victory tells the more interesting story: AMD's wins are often decisive, while Intel's are narrower in some cases but substantial in others.

The largest single delta in the entire comparison belongs to AMD in passmark_integer_math. The Ryzen AI Max+ 392 scores 152414 against Intel's 92150, a 65.4% advantage. That is a commanding lead in a workload that stresses raw integer throughput, and it aligns with the AMD part's 12 cores and 24 threads versus Intel's 14 cores and 14 threads. The absence of hyperthreading on the Intel side appears to cost it heavily in this test.

AMD also dominates passmark_extended_instructions, posting 45666 versus 31212, a 46.3% delta. Similarly, passmark_data_compression shows AMD at 554760 against Intel's 393071, a 41.1% lead. These three results form the core of AMD's case: in dense computational tasks that can use many threads and wide instruction sets, the Ryzen AI Max+ 392 pulls far ahead.

The multithread benchmark reinforces that pattern. AMD scores 45231, Intel scores 38069, a 18.8% delta. The physics test also goes to AMD, 2887 versus 2526, a 14.3% margin. Random string sorting, a memory-latency-sensitive workload, favors AMD as well: 59487 versus 47616, a 24.9% delta.

Intel's counterattack comes in single-thread performance and floating-point math. The Core Ultra 5 245HX scores 4481 in passmark_single_thread against AMD's 3927, a 12.4% advantage. That is a meaningful gap for lightly threaded applications and everyday responsiveness. In passmark_floating_point_math, Intel wins 120735 versus 99548, a 17.5% delta. Intel's architecture clearly handles FP-heavy code more efficiently, despite having fewer threads.

The remaining Intel wins are smaller but notable. In passmark_find_prime_numbers, Intel scores 359 versus 320, a 10.9% delta. In passmark_data_encryption, Intel scores 30022 versus 27784, a 7.5% delta. These are not blowouts, but they show consistency in certain algorithmic patterns.

Looking at aggregate standing, the AMD Ryzen AI Max+ 392 holds a 96th percentile rank among all CPUs in the database, while the Intel Core Ultra 5 245HX sits at the 90th percentile. The average benchmark score tells a similar story: AMD averages 90541, while Intel averages 48287. That large gap in average score is driven by AMD's massive wins in integer math and data compression, which are heavy contributors to the aggregate metric.

AMD's nearest rivals in the database are the Intel Xeon 654 with an average score of 90717 (0.2% ahead of AMD), the AMD Ryzen 9 9955HX at 91199 (0.7% ahead), the Intel Xeon w7-2575X at 88172 (2.7% behind AMD), and the Intel Xeon 6736P at 87864 (3% behind AMD). The Ryzen AI Max+ 392 sits in a performance class that includes workstation Xeon parts, not just mobile chips.

Intel's nearest rivals are notably weaker in average score. The Core Ultra 5 245HX (48287) is essentially tied with the Intel Core Ultra 5 235A (48201, 0.2% behind Intel), the AMD Ryzen AI Max PRO 380 (48171, 0.2% behind), and trails the AMD EPYC 4345P (48470, 0.4% ahead) and Intel Core i5-14600K (48618, 0.7% ahead). The Intel chip's aggregate score places it in a much lower performance tier.

FAQ

Q: Which processor wins more head-to-head benchmarks?

A: The AMD Ryzen AI Max+ 392 wins 6 of the 11 shared tests, while the Intel Core Ultra 5 245HX wins 5.

Q: How large is AMD's biggest winning margin?

A: In passmark_integer_math, AMD scores 152414 against Intel's 92150, a 65.4% advantage.

Q: Where does Intel have its largest advantage?

A: Intel leads by 17.5% in passmark_floating_point_math, scoring 120735 versus AMD's 99548. Intel also leads by 12.4% in passmark_single_thread.

Q: How do the two compare in multithreaded performance?

A: AMD wins passmark_multithread with 45231 against Intel's 38069, an 18.8% margin.

Q: What are the aggregate benchmark scores?

A: AMD averages 90541 and sits at the 96th percentile of all CPUs. Intel averages 48287 and sits at the 90th percentile.

Q: Does Intel have any advantage in core count?

A: Intel has more physical cores (14 versus 12), but AMD has more threads (24 versus 14). Intel's cores do not support simultaneous multithreading in this configuration.

Architecture Differences

The two processors come from fundamentally different design philosophies. AMD's Ryzen AI Max+ 392 uses the Zen 5 architecture under the Strix Halo codename, built on a 4 nm process at TSMC. Intel's Core Ultra 5 245HX uses the Arrow Lake architecture under the Arrow Lake-HX codename, built on a 3 nm process, also at TSMC. Both are mobile parts, but their internal structures diverge sharply.

AMD configures the Ryzen AI Max+ 392 with 12 cores and 24 threads. Each core carries 80 KB of L1 cache, 1 MB of L2 cache, and the chip shares 64 MB of L3 cache. The die is split into two pieces, each measuring 70.6 mm², for a combined die size that reflects the chiplet approach. The large shared L3 is a notable feature: 64 MB is a substantial pool for a mobile processor, and it likely contributes to AMD's strong showing in data compression and integer math.

Intel's Core Ultra 5 245HX uses 14 cores but only 14 threads, meaning no hyperthreading on this SKU. Each core has 192 KB of L1 cache and 3 MB of L2 cache, with 24 MB of shared L3. The Intel die is a monolithic 243 mm² with 17,800 million transistors. The larger per-core L2 cache (3 MB versus 1 MB) may explain Intel's better single-thread score, as more cache per core can reduce memory latency for lightly threaded code.

Memory architecture also differs significantly. AMD uses LPDDR5X with a quad-channel memory bus, delivering 256.0 GB/s of bandwidth. Intel uses DDR5 with a dual-channel bus, delivering 102.4 GB/s. That is a 2.5x difference in theoretical memory bandwidth, which helps explain AMD's dominance in bandwidth-hungry workloads like random string sorting and data compression.

PCIe connectivity differs as well. AMD provides Gen 4 with 16 lanes. Intel provides Gen 5 with 20 lanes. Intel's newer PCIe standard and higher lane count give it an edge for external device bandwidth, though the CPU-only lane counts do not reflect total platform connectivity.

ECC memory support is another differentiator. AMD supports ECC, Intel does not. This positions the AMD part for workstation-adjacent reliability use cases, even within a mobile form factor.

The integrated graphics differ completely. AMD pairs the CPU with Radeon 8060S graphics. Intel uses Arc Xe-LPG Graphics with 48 execution units. The database does not include graphics benchmarks for either part, so a direct comparison of iGPU performance is not possible from this data.

Production status is active for both. AMD's release date is listed as 2026-01-05, while Intel's is 2025-01-12. The multiplier is unlocked on the Intel part, but locked on the AMD part, which matters for overclocking-oriented users.

Specification Differences

The two processors differ in several core specifications. AMD offers 12 cores and 24 threads; Intel offers 14 cores and 14 threads. Base clocks are close: 3.20 GHz for AMD, 3.10 GHz for Intel. Boost clocks are also close: 5.00 GHz for AMD, 5.10 GHz for Intel. Both have a TDP of 55 watts.

Socket and packaging differ: AMD uses AMD Socket FP11, Intel uses Intel BGA 2114. The process nodes differ: AMD is on 4 nm, Intel is on 3 nm. AMD's architecture is Zen 5 with the Strix Halo codename; Intel's is Arrow Lake with the Arrow Lake-HX codename.

Cache configurations are markedly different. AMD provides 80 KB L1 per core, 1 MB L2 per core, and 64 MB shared L3. Intel provides 192 KB L1 per core, 3 MB L2 per core, and 24 MB shared L3. Intel has the larger per-core caches, AMD has the much larger shared L3.

Memory support differs: AMD uses LPDDR5X with a quad-channel bus and 256.0 GB/s bandwidth; Intel uses DDR5 with a dual-channel bus and 102.4 GB/s bandwidth. ECC is supported on AMD, not on Intel.

PCIe generation and lane counts differ: AMD has Gen 4 with 16 lanes, Intel has Gen 5 with 20 lanes. The integrated graphics are different models: Radeon 8060S on AMD, Arc Xe-LPG Graphics 48EU on Intel.

The multiplier is locked on AMD and unlocked on Intel. Intel's transistor count is listed at 17,800 million with a die size of 243 mm²; AMD's die is composed of two 70.6 mm² pieces, with no transistor count listed.

The Verdict

The data points to a clear division of labor. The AMD Ryzen AI Max+ 392 is the stronger processor for multithreaded, bandwidth-intensive, and integer-heavy workloads. Its 65.4% lead in integer math, 46.3% lead in extended instructions, and 41.1% lead in data compression are decisive. The 256.0 GB/s memory bandwidth, quad-channel LPDDR5X support, and 64 MB shared L3 cache give it a structural advantage in tasks that move large amounts of data. The 96th percentile rank and average score of 90541 place it near workstation-class Xeon processors.

The Intel Core Ultra 5 245HX is the better choice for single-threaded performance and floating-point math. Its 12.4% lead in single-thread score and 17.5% lead in floating-point math suggest better per-core efficiency in those patterns. The larger per-core L2 cache (3 MB versus 1 MB) and higher boost clock (5.10 GHz versus 5.00 GHz) support that result. The 90th percentile rank and average score of 48287 put it in a lower overall tier, but the chip still holds its own in specific workloads.

For users who run heavily threaded compilation, data processing, or encryption-adjacent workloads that scale with threads and memory bandwidth, the AMD part is the clear winner. For users who prioritize single-thread responsiveness and floating-point calculations, the Intel part offers a measurable advantage. The Intel chip also has an unlocked multiplier, which may appeal to overclockers despite its lower aggregate performance.

The average benchmark scores are not close: 90541 for AMD versus 48287 for Intel. That gap reflects AMD's dominance in the highest-weight benchmarks. The Intel chip's wins are real but concentrated in fewer, lower-magnitude tests. The database positions the AMD Ryzen AI Max+ 392 in a different performance class entirely, closer to server and workstation silicon than to mainstream mobile parts.