AMD Ryzen AI 7 350 vs Intel Core Ultra 7 165H Comparison

The AMD Ryzen AI 7 350 and Intel Core Ultra 7 165H are evenly matched mobile processors, with the Intel part edging out a narrow 8-to-7 benchmark win total, but the AMD chip winning the more modern and demanding workloads. The data shows a clear split: the Ryzen AI 7 350 dominates in single-threaded and Cinebench R23 multi-threaded tests, while the Core Ultra 7 165H leads in most Passmark multi-threaded and math-specific workloads. The Verdict hinges on workload priority: the AMD chip is the better choice for content creation and single-thread responsiveness, while the Intel chip is preferable for sustained multi-threaded integer and physics tasks. Both CPUs sit at the 84th percentile of all CPUs, and their average benchmark scores are nearly identical (34222 vs 34083), making the choice a matter of specific application performance rather than overall capability.

The Verdict

Choose the AMD Ryzen AI 7 350 if your priority is raw single-core performance and modern multi-threaded rendering. It leads by 11.3% in Cinebench R23 single-core (1958 vs 1759) and by 10.1% in Cinebench R23 multi-core (16014.5 vs 14551). This indicates superior architectural efficiency for both short bursts and sustained rendering tasks. The AMD chip also wins the Passmark single-thread test by 9.3% (3834 vs 3509), making it the better pick for general responsiveness and lightly-threaded applications.

Choose the Intel Core Ultra 7 165H if your workloads involve heavy math, physics simulation, or data encryption. The Intel part wins the Passmark physics test by 25.3% (1806 vs 1349) and floating-point math by 19.7% (66273 vs 53230). It also leads in data encryption by 11.2% (17169 vs 15244) and integer math by 4.2% (89415 vs 85651). For users running scientific or financial applications that leverage these specific instruction patterns, the Intel chip is the data-supported winner.

For a balanced comparison, the deltaPct against each other in their nearestRivals lists is minimal: the AMD chip is 0.4% ahead of Intel in Intel’s own rival list, while Intel is -0.4% behind AMD in AMD’s list. This statistical tie means neither CPU has a definitive overall advantage. The decision should be driven by the specific benchmark wins outlined above, not by overall averages.

Architecture Differences

The two processors are built on fundamentally different architectures and manufacturing processes. The AMD Ryzen AI 7 350 uses the Zen 5 architecture on a 4 nm TSMC process, with the codename Krackan Point. It features 8 cores and 16 threads, with a base clock of 2.00 GHz and a boost clock of 5.00 GHz. The Intel Core Ultra 7 165H uses the Meteor Lake architecture on a 7 nm Intel process, with 16 cores and 22 threads, a base clock of 3.80 GHz, and the same 5.00 GHz boost clock. The Intel chip’s higher base clock reflects its different core design, which relies on more physical cores (16 vs 8) rather than higher single-core efficiency.

Cache configurations differ significantly. The AMD chip has 80 KB of L1 cache per core, 1 MB of L2 per core, and 8 MB of L3 cache. The Intel chip has 112 KB of L1 per core, 2 MB of L2 per core, and 24 MB of shared L3 cache. This larger cache pool on Intel likely contributes to its wins in data compression and random string sorting, where larger working sets benefit from more L3 capacity. The AMD chip’s smaller L3 cache is offset by its superior single-core architecture, which shows in the 14.4% lead in Cinebench R15 single-core (294 vs 257).

Both support DDR5 and LPDDR5X memory with dual-channel buses and identical 89.6 GB/s memory bandwidth. They also share the same 28 W TDP and do not support ECC memory. The AMD chip uses PCIe Gen 4 with 16 CPU lanes, while the Intel chip uses PCIe Gen 5 with 8 CPU lanes. The integrated graphics differ: AMD pairs the Radeon 860M, while Intel uses Arc Xe-LPG 128EU. Neither chip has an unlocked multiplier, and both are currently in active production.

Where Each One Wins

The AMD Ryzen AI 7 350 wins in all single-threaded benchmarks and in the Cinebench R23 multi-core test. Its 14.4% lead in Cinebench R15 single-core (294 vs 257) and 11.3% lead in R23 single-core (1958 vs 1759) are the largest margins in the comparison. The Passmark single-thread test shows a 9.3% lead (3834 vs 3509). This CPU is the clear choice for tasks like web browsing, office productivity, and legacy single-threaded applications. For multi-core rendering, the Cinebench R23 result (16014.5 vs 14551) shows AMD’s architecture handles modern multi-threaded workloads better than Intel’s, despite having half the physical cores. The AMD chip also wins extended instructions by 25.3% (21678 vs 17299) and data compression by 4.4% (304089 vs 291161), making it strong for compression utilities and vectorized code.

The Intel Core Ultra 7 165H wins in most Passmark multi-threaded tests, leveraging its higher core count. The physics test shows a 25.3% lead (1806 vs 1349), floating-point math a 19.7% lead (66273 vs 53230), and data encryption an 11.2% lead (17169 vs 15244). The integer math win is 4.2% (89415 vs 85651), and the multithread test shows a 3.8% lead (25933 vs 24935). The random string sorting win is marginal at 0.6% (33473 vs 33266). These wins indicate Intel’s advantage in parallel compute tasks that scale with core count, particularly in physics engines, encryption algorithms, and floating-point-heavy workloads. The Cinebench R15 multi-core test also goes to Intel by 6.2% (2641 vs 2477), showing that older multi-threaded benchmarks favor Intel’s core count.

FAQ

Q: Which CPU has the higher single-core performance?

A: The AMD Ryzen AI 7 350 wins every single-threaded benchmark in the data: Cinebench R15 single-core by 14.4% (294 vs 257), Cinebench R23 single-core by 11.3% (1958 vs 1759), and Passmark single-thread by 9.3% (3834 vs 3509).

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

A: The AMD Ryzen AI 7 350 wins the Cinebench R23 multi-core test by 10.1% (16014.5 vs 14551), but Intel wins Cinebench R15 multi-core by 6.2% (2641 vs 2477) and Passmark multithread by 3.8% (25933 vs 24935). The winner depends on the specific rendering engine.

Q: Which CPU has more cores?

A: The Intel Core Ultra 7 165H has 16 cores and 22 threads, while the AMD Ryzen AI 7 350 has 8 cores and 16 threads. Despite fewer cores, AMD wins the modern Cinebench R23 multi-core test.

Q: Are their overall benchmark scores similar?

A: Yes. The AMD chip has an average benchmark score of 34222, and Intel has 34083, a difference of only 0.4%. Both CPUs are ranked at the 84th percentile of all CPUs.

Q: Which CPU is better for physics simulations?

A: The Intel Core Ultra 7 165H wins the Passmark physics test by 25.3% (1806 vs 1349), making it the clear choice for physics-based workloads.

Q: What is the TDP of both CPUs?

A: Both the AMD Ryzen AI 7 350 and Intel Core Ultra 7 165H have a TDP of 28 W, meaning they are designed for similar thermal envelopes in mobile devices.

Head-to-Head Benchmarks

The most significant win for the AMD Ryzen AI 7 350 is in Passmark extended instructions, where it beats Intel by 25.3% (21678 vs 17299). This indicates a major advantage in SIMD and vectorized workloads, which are common in modern media encoding and scientific computing. The second-largest AMD win is in Cinebench R15 single-core, at 14.4% (294 vs 257), demonstrating the Zen 5 architecture’s superior per-core execution. The Cinebench R23 multi-core win at 10.1% (16014.5 vs 14551) is the most important for content creators, showing that AMD’s 8 cores outperform Intel’s 16 in this modern rendering benchmark. The 9.3% Passmark single-thread win (3834 vs 3509) further confirms AMD’s single-core dominance.

The Intel Core Ultra 7 165H’s largest win is in Passmark physics, at 25.3% (1806 vs 1349). This is a massive margin that makes Intel the obvious choice for physics-based simulation and gaming physics. The floating-point math win is also substantial at 19.7% (66273 vs 53230), indicating Intel’s advantage in heavy mathematical computations. The data encryption win of 11.2% (17169 vs 15244) is notable for security-related workloads. The find prime numbers test shows Intel winning by 32.2% (118 vs 80), which is the largest single-test margin in the entire comparison, though it represents a narrow absolute gap. Intel also wins Cinebench R15 multi-core by 6.2% (2641 vs 2477), Passmark integer math by 4.2% (89415 vs 85651), and Passmark multithread by 3.8% (25933 vs 24935). The random string sorting test is essentially a tie, with Intel winning by only 0.6% (33473 vs 33266).

Specification Differences

The most striking specification difference is core count: Intel has 16 cores and 22 threads, while AMD has 8 cores and 16 threads. This explains Intel’s wins in core-scalable workloads, but AMD’s architecture compensates with higher single-core efficiency. The base clocks differ significantly, with Intel at 3.80 GHz versus AMD’s 2.00 GHz, while both boost to 5.00 GHz. The process node differs: AMD uses 4 nm TSMC, Intel uses 7 nm Intel. Cache sizes vary, with Intel having larger per-core L1 (112 KB vs 80 KB) and L2 (2 MB vs 1 MB), plus a much larger shared L3 (24 MB vs 8 MB). The socket is different (AMD Socket FP8 vs Intel BGA 2049). PCIe support differs: AMD uses Gen 4 with 16 lanes, Intel uses Gen 5 with 8 lanes. The integrated GPUs are different (Radeon 860M vs Arc Xe-LPG 128EU). The release dates are also different: AMD launched on 2025-01-05, Intel on 2023-12-13. Intel has a launch MSRP of $460, while AMD has no listed launch MSRP. Both have a 28 W TDP, dual-channel memory buses, and 89.6 GB/s memory bandwidth. The Intel die size is not listed, while AMD’s is 195 mm².