AMD Ryzen AI Max PRO 390 vs Intel Core 9 273PQE Comparison

The Intel Core 9 273PQE and the AMD Ryzen AI Max PRO 390 occupy the same statistical neighborhood in the database, both sitting in the 93rd percentile of all recorded CPUs, yet they get there by completely different routes. The Intel part is a 125 W desktop processor that converts raw clock speed into decisive single-thread and multi-thread wins, taking 11 of 15 head-to-head benchmarks. The AMD part is a 55 W mobile design built on a denser process node, and it wins the tests that reward memory bandwidth and specialized instruction throughput. The overall gap in average benchmark score is modest, 66099 for Intel versus 63765 for AMD, but the per-test distribution tells a more polarized story than that aggregate suggests.

Where Each One Wins

The Core 9 273PQE dominates wherever per-core speed is the bottleneck. Its Cinebench R23 single-core score of 5532 against 1976 for the AMD chip is a 180% advantage, and the R23 multi-core result of 39190 against 24828 extends that lead by 57.8% once all 24 threads are engaged. Passmark single-thread, integer math, floating point math, data compression, and data encryption all fall to Intel by margins ranging from roughly 11% to 33%. If a workload is latency-sensitive, lightly threaded, or a classic rendering-style task, the recorded data leaves no ambiguity: the Intel chip wins.

The Ryzen AI Max PRO 390 wins four tests, and they cluster around a clear theme. It takes passmark find prime numbers at 323 versus 198, a 38.7% lead, extended instructions at 40888 versus 38743, physics at 2773 versus 2754, and random string sorting at 55248 versus 53167. These are workloads that benefit from memory subsystem behavior and specific instruction sets rather than pure frequency, and they align with the AMD chip's 256.0 GB/s of memory bandwidth on a quad-channel LPDDR5X bus. For bandwidth-hungry integer and vector workloads, and for physics simulation, the data shows the Strix Halo design holding its own despite operating at less than half the TDP.

FAQ

Q: Which CPU is faster overall?

A: The Intel Core 9 273PQE. It wins 11 of 15 head-to-head benchmarks and posts the higher average benchmark score, 66099 versus 63765, a gap of about 3.7%.

Q: How big is the single-core difference?

A: Large. The Core 9 273PQE scores 5532 in Cinebench R23 single-core against 1976 for the Ryzen AI Max PRO 390, a 180% advantage, and leads passmark single-thread 4573 to 3967, or 15.3%.

Q: Does the AMD chip win anything?

A: Yes, four tests: find prime numbers (38.7% ahead), extended instructions (5.2% ahead), physics (0.7% ahead), and random string sorting (3.8% ahead).

Q: Are multi-core results close?

A: In Cinebench R15 multi-core, essentially yes: 3950 versus 3918, only 0.8% apart. In Cinebench R23 multi-core, no: Intel leads 39190 to 24828, a 57.8% gap. Passmark multithread sits in between at 46107 versus 42912, 7.4% apart.

Q: How do their rival groups compare?

A: The Core 9 273PQE is statistically bracketed by the Intel Core Ultra 5 250KF Plus (-0.1%), AMD Ryzen 9 7950X3D (+0.3%), Intel Core Ultra 5 250K Plus (-1.1%), and AMD EPYC 4465P (-1.2%). The Ryzen AI Max PRO 390 sits near the Intel Core i9-13900KS (-0.4%), AMD Ryzen AI 7 450G (+0.7%), AMD EPYC 7343 (-0.7%), and Intel Core Ultra 7 265HX (+0.9%). Both sit in the 93rd percentile of all CPUs in the database.

Q: What is the launch MSRP of the Core 9 273PQE?

A: Its launch MSRP is $589. No launch MSRP figure is recorded for the Ryzen AI Max PRO 390.

Head-to-Head Benchmarks

The single largest Intel win is Cinebench R23 single-core: 5532 against 1976, a 180% margin that reflects the Core 9 273PQE's 5.90 GHz boost clock against the AMD chip's 5.00 GHz. Cinebench R15 single-core shows the same pattern at 557 versus 303, an 83.8% gap. These are the defining results of the comparison: when one core does the work, the Intel processor is in a different performance class.

Sustained multi-core rendering splits by generation of the test. In the older Cinebench R15 multi-core run the two are nearly tied, 3950 versus 3918, a 0.8% Intel edge. In Cinebench R23 the gap explodes to 57.8% in Intel's favor, 39190 to 24828, suggesting the newer, longer, more AVX-intensive workload plays to the Bartlett Lake design's strengths. Passmark multithread lands at 46107 versus 42912, a 7.4% Intel win.

The rest of the Passmark suite favors Intel by moderate margins: floating point math 125546 to 94666 (+32.6%), data compression 585752 to 506170 (+15.7%), single-thread 4573 to 3967 (+15.3%), data encryption 29636 to 26027 (+13.9%), and integer math 164629 to 148508 (+10.9%).

AMD's wins are fewer but not noise-level in one case. Find prime numbers is its standout, 323 versus 198, a 38.7% lead, indicating strong per-cycle throughput in that integer workload despite the lower clocks. Extended instructions goes to AMD at 40888 versus 38743 (+5.2%), random string sorting at 55248 versus 53167 (+3.8%), and physics at 2773 versus 2754 (+0.7%). The physics result is effectively a tie, but it is notable that the 55 W mobile part matches a 125 W desktop chip at all in a simulation workload.

Specification Differences

The two processors share core count (12 cores, 24 threads) and per-core L1 cache (80 KB), but diverge sharply elsewhere. The Core 9 273PQE is a desktop part on Intel Socket 1700 with a 125 W TDP, a 3.40 GHz base clock, and a 5.90 GHz boost clock. The Ryzen AI Max PRO 390 is a mobile part on AMD Socket FP11 with a 55 W TDP, a 3.20 GHz base clock, and a 5.00 GHz boost clock.

Memory is where the contrast is sharpest. Intel supports DDR4 and DDR5 on a dual-channel bus at 89.6 GB/s of bandwidth. AMD supports LPDDR5X on a quad-channel bus delivering 256.0 GB/s, nearly triple the bandwidth, which explains its wins in bandwidth-sensitive Passmark tests. Both support ECC memory.

Platform connectivity also differs: Intel offers PCIe Gen 5 with 16 CPU lanes, while AMD provides PCIe Gen 4 with 16 CPU lanes. The integrated graphics differ, with Intel pairing UHD Graphics 770 and AMD pairing the Radeon 8050S. The Intel part carries a launch MSRP of $589; no launch MSRP is recorded for the AMD part. Both are active products with locked multipliers.

Architecture Differences

The Core 9 273PQE is built on Intel's 10 nm process at Intel's own foundry, codenamed Bartlett Lake. Its cache hierarchy emphasizes large per-core L2, at 2 MB per core, with 36 MB of shared L3. That fat L2 suits a high-clock design that wants data close to fast cores.

The Ryzen AI Max PRO 390 is a Zen 5 design codenamed Strix Halo, manufactured by TSMC on a 4 nm node. It halves per-core L2 to 1 MB but doubles shared L3 to 64 MB, a split consistent with a chip designed to feed a wide, bandwidth-rich memory system rather than to maximize per-core frequency. The newer node is also what allows the design to deliver competitive multi-core results at a 55 W TDP, well under half of the Intel chip's 125 W envelope.

In short: Intel's architecture bets on clock speed and per-core cache, AMD's bets on process density, a large shared cache, and memory bandwidth. The benchmark data validates both bets in their respective territories.

The Verdict

For desktop users whose workloads are single-threaded or rendering-centric, the data points unambiguously to the Core 9 273PQE. A 180% single-core lead in Cinebench R23, a 57.8% multi-core lead in the same test, and wins across nearly every general-purpose Passmark metric make it the performance pick of the pair, provided the 125 W TDP and Socket 1700 platform fit the build.

The Ryzen AI Max PRO 390 is the choice where efficiency, memory bandwidth, and platform integration matter more than peak per-core speed. It delivers comparable Cinebench R15 multi-core output (within 0.8%), wins prime number, extended instruction, string sorting, and physics workloads, and does so at 55 W on a quad-channel 256.0 GB/s memory subsystem with a Radeon 8050S integrated GPU, all in a mobile package on Socket FP11.

Both processors rank in the 93rd percentile of the database, and both are bracketed by strong rival groups, Intel near the Ryzen 9 7950X3D and AMD near the Core i9-13900KS. The recorded data frames this not as a close call, but as a split decision along clear lines: raw speed for Intel, bandwidth and efficiency for AMD.