AMD Ryzen AI Max PRO 390 vs Intel Core Ultra 7 265F Comparison

The Intel Core Ultra 7 265F and AMD Ryzen AI Max PRO 390 are both 93rd-percentile performers, yet they are fundamentally different processors. The Intel part is a 20-thread desktop chip built on a 3 nm TSMC process, while the AMD part is a 12-core mobile chip with 24 threads and a large shared cache. Benchmark data shows the Intel chip wins 13 of 15 head-to-head tests, but the AMD chip takes two meaningful victories in integer math and extended instructions. The Intel Core Ultra 7 265F is the clear compute leader in most threaded and single-threaded workloads, but the AMD Ryzen AI Max PRO 390 counters with a higher memory bandwidth and an integrated GPU, making it a different kind of animal entirely.

Where Each One Wins

The Intel Core Ultra 7 265F dominates in almost every category that stresses raw CPU throughput. Its largest advantages come in single-core performance, where it leads by 97% in Cinebench R15 single-core and by 199.9% in Cinebench R23 single-core. This makes it the obvious choice for lightly-threaded applications, legacy software, and any workload that relies on a single fast core. The data also shows a massive 83.7% lead in floating-point math, which points to a strong showing in scientific computing, physics simulation, and rendering tasks that use FP32 or FP64 instructions. The Intel chip also wins in multi-core tests, with a 15.1% lead in PassMark multithread and a 69.1% lead in Cinebench R23 multi-core, cementing its position for video encoding, 3D rendering, and heavy compilation work.

The AMD Ryzen AI Max PRO 390, despite losing the overall count, wins where its architecture is strongest. It takes the PassMark integer math test by 7%, which suggests an edge in certain database, cryptography, or compression workloads that rely on integer operations. It also wins the extended instructions test by 4%, indicating better utilization of AVX-512 or similar advanced instruction sets. The AMD chip’s 64 MB of shared L3 cache—more than double the Intel’s 30 MB—likely contributes to these wins, especially in data-heavy loops. Beyond compute, the AMD part has a Radeon 8050S integrated GPU, while the Intel chip has no integrated graphics, making the AMD the pick for a compact system without a discrete GPU. The AMD’s quad-channel LPDDR5X memory bus also delivers 256.0 GB/s bandwidth, versus the Intel’s dual-channel DDR5 at 102.4 GB/s, which matters for memory-bound tasks.

FAQ

Q: Which processor is faster in single-core benchmarks?

A: The Intel Core Ultra 7 265F is dramatically faster. It scores 597 in Cinebench R15 single-core and 5926 in Cinebench R23 single-core, versus 303 and 1976 for the AMD Ryzen AI Max PRO 390. The deltas are 97% and 199.9%, respectively.

Q: Does the AMD chip win any benchmark at all?

A: Yes, the AMD Ryzen AI Max PRO 390 wins two tests: PassMark integer math (148508 vs 138078, a 7% lead) and PassMark extended instructions (40888 vs 39235, a 4% lead). The Intel chip wins the other 13 head-to-head tests.

Q: Which CPU is better for multi-threaded rendering?

A: The Intel Core Ultra 7 265F. It scores 41980 in Cinebench R23 multi-core compared to 24828 for the AMD, a 69.1% advantage. It also leads PassMark multithread by 15.1% (49410 vs 42912).

Q: What about memory bandwidth? Does that favor one?

A: Yes, the AMD Ryzen AI Max PRO 390 has a quad-channel LPDDR5X memory bus with 256.0 GB/s bandwidth, while the Intel Core Ultra 7 265F uses dual-channel DDR5 at 102.4 GB/s. This gives the AMD a 2.5x theoretical bandwidth advantage, which can help in memory-intensive tasks.

Q: Are these chips comparable in overall score?

A: The average benchmark scores are close. The Intel Core Ultra 7 265F has an average benchmark score of 64438, while the AMD Ryzen AI Max PRO 390 scores 63765. Both sit at the 93rd percentile of all CPUs.

Q: Does the AMD chip have integrated graphics?

A: Yes, the AMD Ryzen AI Max PRO 390 includes a Radeon 8050S iGPU. The Intel Core Ultra 7 265F has no integrated graphics, meaning it requires a separate discrete GPU to output video.

Head-to-Head Benchmarks

The largest single win for the Intel Core Ultra 7 265F is in Cinebench R23 single-core, where it scores 5926 against the AMD’s 1976—a staggering 199.9% delta. That is not a marginal lead; it is a generational gap in IPC and clock speed. The Intel chip boosts to 5.30 GHz versus 5.00 GHz for the AMD, and its 2.40 GHz base clock is lower, but the raw single-thread performance suggests the Arrow Lake architecture is far more efficient per clock in this test. Similarly, Cinebench R15 single-core shows a 97% lead (597 vs 303), reinforcing that the Intel part is the one for single-threaded software.

In multi-core, the Intel chip’s advantage is also pronounced, though less extreme. Cinebench R23 multi-core shows 41980 for Intel and 24828 for AMD, a 69.1% lead. The Intel chip has 20 cores and 20 threads, while the AMD has 12 cores and 24 threads. The AMD’s extra threads do not compensate for the core count deficit, likely due to the Intel’s higher boost clocks and superior memory latency. The PassMark multithread test tells a similar story: Intel wins 49410 to 42912, a 15.1% lead. The AMD’s wins are narrower but real. In integer math, the AMD scores 148508 versus 138078, a 7% lead. In extended instructions, it scores 40888 versus 39235, a 4% lead. These are not blowouts, but they are consistent, suggesting the Zen 5 architecture handles specific instruction patterns better than Arrow Lake.

Other notable Intel wins include data encryption (39468 vs 26027, a 51.6% lead), floating-point math (173855 vs 94666, an 83.7% lead), and find prime numbers (416 vs 323, a 28.8% lead). The Intel chip also wins in physics (3172 vs 2773, 14.4%), random string sorting (62439 vs 55248, 13%), and single-thread PassMark (4750 vs 3967, 19.7%). The only near-tie is data compression, where Intel scores 507018 and AMD scores 506170, a 0.2% difference that is effectively noise.

Specification Differences

The core and thread counts differ significantly. The Intel Core Ultra 7 265F has 20 cores and 20 threads, while the AMD Ryzen AI Max PRO 390 has 12 cores and 24 threads. The Intel chip has a base clock of 2.40 GHz and a boost clock of 5.30 GHz; the AMD runs at 3.20 GHz base and 5.00 GHz boost. The Intel part is a desktop chip on Intel Socket 1851, whereas the AMD is a mobile chip on AMD Socket FP11. The Intel chip has a TDP of 65 watts, and the AMD has a TDP of 55 watts.

Memory support also diverges. The Intel chip uses DDR5 over a dual-channel bus, yielding 102.4 GB/s bandwidth. The AMD chip uses LPDDR5X over a quad-channel bus, yielding 256.0 GB/s. The AMD also supports ECC memory, while the Intel does not. PCIe lanes differ: the Intel offers Gen 5 with 20 lanes (CPU only), and the AMD offers Gen 4 with 16 lanes (CPU only). The Intel has no integrated graphics, while the AMD includes a Radeon 8050S. The Intel chip’s launch MSRP is $379; the AMD has no launch MSRP listed.

Architecture Differences

The Intel Core Ultra 7 265F is built on Arrow Lake-S, part of the Core Ultra Series 2, using TSMC’s 3 nm process node. It packs 17,800 million transistors on a 243 mm² die. Its cache layout includes 192 KB of L1 per core, 3 MB of L2 per core, and 30 MB of shared L3 cache. The Intel architecture is a desktop-first design, with a high transistor count and a small process node, but it lacks a v-cache and has no integrated graphics. The production status is active, and it was released on January 6, 2025.

The AMD Ryzen AI Max PRO 390 is based on Zen 5 and codenamed Strix Halo. It uses a 4 nm TSMC process, but the data does not list a transistor count or die size. Its cache configuration shows 80 KB of L1 per core, 1 MB of L2 per core, and a much larger 64 MB of shared L3 cache. The AMD chip is a mobile part, designed for laptops and compact systems, and it includes a Radeon 8050S iGPU. It supports ECC memory and uses a quad-channel LPDDR5X interface. The AMD chip was released one day earlier, on January 5, 2025. The larger L3 cache on the AMD part, combined with the higher memory bandwidth, suggests an architecture tuned for data locality and memory throughput, while the Intel part relies on raw clock speed and core count to win benchmarks.