AMD Ryzen AI Max PRO 390 vs Intel Core Ultra 5 336H Comparison

Head-to-Head Benchmarks

The recorded data shows a clear overall winner in the AMD Ryzen AI Max PRO 390, which takes 11 of the 15 head-to-head benchmark comparisons. The margin of victory is often substantial. In Cinebench R15 multi-core, the AMD part scores 3918 against 2418 for the Intel Core Ultra 5 336H, a 62% advantage. The PassMark integer math test shows the largest gap, with AMD at 148508 versus Intel's 62070, a delta of 139.3%. Data compression also heavily favors AMD, with a score of 506170 compared to 280340, an 80.6% lead. Extended instructions follow a similar pattern, 40888 versus 24542, a 66.6% difference. Random string sorting goes to AMD at 55248 versus 34402, a 60.6% win. The multi-threaded PassMark score shows AMD ahead at 42912 against 28545, a 50.3% margin. Data encryption, floating point math, prime number finding, and physics all add to the AMD tally, with deltas of 22.2%, 14.9%, 8%, and 3.4% respectively. Even the Cinebench R23 multi-core test, the closest multi-threaded result, goes to AMD at 24828 versus 23991, a 3.5% edge.

The Intel Core Ultra 5 336H does secure four wins, and they are concentrated in single-threaded workloads. The most dramatic victory is in Cinebench R23 single-core, where Intel scores 3387 against AMD's 1976, a 41.7% advantage. The Cinebench R15 single-core test also goes to Intel, 341 versus 303, an 11.1% lead. In the PassMark single-thread test, Intel edges out AMD with 4013 versus 3967, a slim 1.1% margin. That same score appears twice in the data, confirming the result. These single-core wins show a distinct strength for the Intel architecture, but they are outweighed by the sheer number and magnitude of the AMD multi-threaded victories.

Architecture Differences

The two processors are built on fundamentally different foundations. The AMD Ryzen AI Max PRO 390 uses the Zen 5 architecture, codenamed Strix Halo, and is manufactured on a 4 nm process by TSMC. It features 12 cores and 24 threads, enabling simultaneous multi-threading. The Intel Core Ultra 5 336H uses the Panther Lake architecture, also its codename, and is built on a 3 nm process by Intel. It has 16 cores but only 16 threads, meaning it lacks hyper-threading.

Cache configurations differ substantially. The AMD chip allocates 80 KB of L1 per core and 1 MB of L2 per core, with a large 64 MB shared L3 cache. The Intel part has 192 KB of L1 per core and 2.5 MB of L2 per core, but only 18 MB of shared L3. This gives AMD a 64 MB L3 pool versus Intel's 18 MB, a significant difference for workloads that benefit from large shared caches. Memory support also diverges: AMD uses LPDDR5X with a quad-channel memory bus, delivering 256.0 GB/s of bandwidth. Intel supports both DDR5 and LPDDR5X, but through a dual-channel bus, capping bandwidth at 115.2 GB/s. AMD also supports ECC memory, while Intel does not. PCIe connectivity differs, with AMD offering Gen 4 across 16 lanes (CPU only) and Intel providing Gen 5 across 12 lanes (CPU only). The integrated graphics are the Radeon 8050S on the AMD side and Intel Xe3 Graphics on the Intel side. The sockets are incompatible, with AMD using Socket FP11 and Intel using BGA 2540. The AMD part has a 55 W TDP, while the Intel part is rated at 25 W. Clock speeds also tell a story: AMD has a base clock of 3.20 GHz and a boost of 5.00 GHz, while Intel has a lower base of 1.90 GHz but a boost of 4.60 GHz. The release dates place AMD in January 2025 and Intel in January 2026.

Where Each One Wins

The benchmark data splits the processors into two distinct usage profiles. AMD dominates in every multi-threaded and throughput-oriented test. The integer math score of 148508 is more than double Intel's 62070, indicating strong performance in general computation, compilation, and data processing. Data compression at 506170 versus 280340 shows AMD handling archiving and file management tasks with ease. Extended instructions, scoring 40888 against 24542, point to AMD's strength in SIMD and vectorized workloads, common in scientific computing and media encoding. The multi-thread PassMark score of 42912 versus 28545 confirms AMD's advantage in parallel tasks. Even the Cinebench R23 multi-core result, while closer, favors AMD, meaning rendering and 3D workloads lean AMD. The physics test, a proxy for simulation and gaming physics calculations, goes to AMD at 2773 versus 2682, a narrow but consistent win.

Intel's wins are exclusively in single-threaded performance, which matters for lightly threaded applications, legacy software, and certain games that rely on a single fast core. The Cinebench R23 single-core score of 3387 is a standout, showing a 41.7% lead over AMD. This is a major gap and suggests that tasks like spreadsheet recalculation, web browsing, and older game engines will feel snappier on the Intel chip. The PassMark single-thread score of 4013 versus 3967 is a narrow margin, but it confirms the trend. For workloads that cannot use multiple cores, the Intel Core Ultra 5 336H is the clear choice. For anything that can scale across cores, the AMD Ryzen AI Max PRO 390 is decisively ahead.

The Verdict

The data presents a straightforward choice based on workload priorities. The AMD Ryzen AI Max PRO 390 should be selected for users whose tasks involve heavy multi-threading, large data sets, content creation, or any parallel processing. Its 64 MB L3 cache and quad-channel memory bandwidth give it a structural advantage in feeding many cores. The 11 benchmark wins, including dominant margins in integer math and data compression, make it the superior processor for productivity and creative work. The Cinebench R23 multi-core score of 24828, while only 3.5% ahead of Intel, still confirms that AMD leads in rendering workloads.

The Intel Core Ultra 5 336H should be chosen when single-threaded performance is the priority. Its 41.7% lead in Cinebench R23 single-core is a massive gap, and it also wins in Cinebench R15 single-core and PassMark single-thread. This makes it suitable for applications that are poorly optimized for multi-threading, where the higher boost clock of 4.60 GHz and the Panther Lake architecture deliver tangible benefits. The lower 25 W TDP also suggests it may fit into more power-constrained designs, though the data does not include power consumption measurements. The Intel part is not without merit, but its four wins come in a narrow category. The overall average benchmark score supports this split: AMD sits at 63765, placing it in the 93rd percentile of all CPUs, while Intel scores 34485, placing it in the 84th percentile. The nearest rivals for AMD include the Intel Core i9-13900KS with a delta of -0.4%, meaning AMD is essentially at parity with that flagship desktop chip. Intel's nearest rival is the Intel Core i7-13700HX, with a delta of -0.2%, indicating it performs similarly to a previous-generation high-end mobile chip. For a balanced mobile platform, the AMD Ryzen AI Max PRO 390 is the stronger overall processor based on the recorded measurements.

FAQ

Q: How many benchmark wins does each processor have?

A: The AMD Ryzen AI Max PRO 390 wins 11 of the 15 head-to-head comparisons, while the Intel Core Ultra 5 336H wins 4.

Q: What is the largest margin of victory in the head-to-head tests?

A: The largest margin is in PassMark integer math, where AMD scores 148508 versus Intel's 62070, a delta of 139.3%.

Q: Which processor has a higher boost clock?

A: The AMD Ryzen AI Max PRO 390 has a boost clock of 5.00 GHz, while the Intel Core Ultra 5 336H has a boost clock of 4.60 GHz.

Q: What is the L3 cache size difference?

A: AMD has 64 MB of shared L3 cache, while Intel has 18 MB of shared L3 cache.

Q: Which processor supports ECC memory?

A: The AMD Ryzen AI Max PRO 390 supports ECC memory, while the Intel Core Ultra 5 336H does not.

Q: What is the memory bandwidth for each?

A: AMD provides 256.0 GB/s over a quad-channel LPDDR5X bus, while Intel provides 115.2 GB/s over a dual-channel DDR5/LPDDR5X bus.

Specification Differences

| Specification | AMD Ryzen AI Max PRO 390 | Intel Core Ultra 5 336H |

|---|---|---|

| Cores | 12 | 16 |

| Threads | 24 | 16 |

| Base Clock | 3.20 GHz | 1.90 GHz |

| Boost Clock | 5.00 GHz | 4.60 GHz |

| TDP | 55 W | 25 W |

| Socket | AMD Socket FP11 | Intel BGA 2540 |

| Architecture | Zen 5 | Panther Lake |

| Process Node | 4 nm | 3 nm |

| Foundry | TSMC | Intel |

| L1 Cache | 80 KB (per core) | 192 KB (per core) |

| L2 Cache | 1 MB (per core) | 2.5 MB (per core) |

| L3 Cache | 64 MB (shared) | 18 MB (shared) |

| Memory Support | LPDDR5X | DDR5, LPDDR5X |

| Memory Bus | Quad-channel | Dual-channel |

| Memory Bandwidth | 256.0 GB/s | 115.2 GB/s |

| ECC Memory | Yes | No |

| PCIe | Gen 4, 16 Lanes (CPU only) | Gen 5, 12 Lanes (CPU only) |

| Integrated Graphics | Radeon 8050S | Intel Xe3 Graphics |

| Release Date | 2025-01-05 | 2026-01-04 |

| Part Number | 100-000001421 | SA4RD |