AMD Ryzen AI 9 365 vs Intel Core i9-13905H Comparison

Head-to-Head Benchmarks

The benchmark data splits almost evenly between these two mobile processors, with the Intel Core i9-13905H taking 8 wins and the AMD Ryzen AI 9 365 taking 7. The margins, however, tell a more nuanced story than the win count suggests. Intel’s biggest victories are decisive, while AMD’s are mostly narrow.

In Cinebench R23 multi-core, the Intel part scores 20034 against AMD’s 18698, a 7.1% lead that reflects the raw throughput advantage of its extra cores. The gap widens further in PassMark floating-point math, where Intel posts 73434 versus 62802 – a 16.9% margin that makes it the clear choice for numerically heavy workloads. Physics simulation is even more lopsided: Intel scores 2054 against 1704, a 20.5% advantage that is the largest single delta in the entire comparison. Data encryption also favors Intel by 11.8% (20464 vs 18297), and prime number finding shows a smaller but still positive 4.3% edge (122 vs 117). In Cinebench R15 multi-core, the two are nearly tied at 2857 vs 2842, with Intel ahead by just 0.5%.

AMD’s wins are concentrated in single-threaded and specialized instruction workloads, but they are mostly modest. The standout is Cinebench R15 single-core, where AMD scores 303 against Intel’s 246 – an 18.8% blowout that is the second-largest delta in the dataset. That result is anomalous compared to the other single-core tests, so it should be treated with some caution. In Cinebench R23 single-core, Intel actually wins 2020 vs 1992, a 1.4% edge. AMD takes PassMark single-thread by 3.6% (3841 vs 3703), extended instructions by 15.2% (25113 vs 21297), and random string sorting by 5% (39447 vs 37460). Data compression and integer math are effectively dead heats, with AMD ahead by 0.1% in both (354510 vs 354330 and 101831 vs 101716 respectively).

Overall averages reflect this split. Intel’s average benchmark score is 40313, AMD’s is 40048, a difference of 0.7%. Both processors sit at the 87th percentile against all CPUs, meaning they land in the same performance tier despite their different design philosophies.

Architecture Differences

The two chips take fundamentally different approaches to mobile computing. Intel’s Core i9-13905H is built on Raptor Lake architecture using a 10nm process at Intel’s own foundry. It packs 14 cores and 20 threads, with a base clock of 2.60 GHz and a boost clock of 5.40 GHz. The die measures 257 mm². AMD’s Ryzen AI 9 365 uses Zen 5 architecture on a 4nm process fabricated by TSMC, with 10 cores and 20 threads. Its base clock is 2.00 GHz and boost reaches 5.00 GHz, on a slightly smaller 233 mm² die.

The core count difference is significant: Intel has 4 more physical cores, though both chips expose 20 threads. Cache layouts also differ. Both allocate 80 KB of L1 per core, but Intel provides 2 MB of L2 per core versus AMD’s 1 MB per core. Intel’s L3 is 24 MB shared, while AMD’s is 16 MB. That extra cache is part of why Intel maintains a multi-core edge despite AMD’s more modern process node.

Memory support diverges as well. Intel supports both DDR4 and DDR5 in dual-channel configuration, while AMD supports DDR5 and LPDDR5X. AMD lists a memory bandwidth of 89.6 GB/s, a figure Intel does not provide. PCIe connectivity differs: Intel offers Gen 5 with 8 CPU lanes, AMD offers Gen 4 with 16 CPU lanes. The integrated graphics are different tiers – Intel uses Iris Xe Graphics with 96 execution units, AMD uses the Radeon 880M. Power targets also diverge, with Intel rated at 45W TDP and AMD at 28W TDP.

The release dates are a generation apart: Intel launched on 2023-01-03, AMD on 2024-06-30. Intel’s part number is SRMHU and its launch MSRP is $697; AMD’s part number is 100-000001530 with no launch MSRP listed. Both are active production parts for the mobile market, neither has an unlocked multiplier, and both lack ECC memory support.

The Verdict

The data paints a clear picture for different use cases. If you prioritize multi-core throughput, encryption performance, floating-point math, or physics simulation, the Intel Core i9-13905H is the right choice. Its 7.1% lead in Cinebench R23 multi-core, 16.9% edge in floating-point math, and 20.5% advantage in physics are substantial enough to matter in real workloads. The 11.8% encryption win also matters for anyone handling sensitive data regularly.

If your work leans on single-threaded responsiveness, extended instruction sets, or string manipulation, the AMD Ryzen AI 9 365 has the edge. Its 15.2% lead in extended instructions and 5% advantage in random string sorting are meaningful for developers and data analysts. The 3.6% single-thread win in PassMark, combined with its lower 28W TDP, makes it the more efficient choice for sustained light workloads.

The overall average scores are close enough that neither chip is a universal winner – 40313 vs 40048 is a 0.7% difference. But the distribution of wins matters more than the average. Intel’s victories are often larger in magnitude, which suggests it handles peak multi-threaded loads better. AMD’s wins are mostly narrow but consistent in the categories it dominates. For a builder who knows their primary applications, the choice is straightforward: Intel for compute-heavy parallel tasks, AMD for instruction diversity and efficiency.

Specification Differences

| Field | Intel Core i9-13905H | AMD Ryzen AI 9 365 |

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

| Cores | 14 | 10 |

| Base Clock | 2.60 GHz | 2.00 GHz |

| Boost Clock | 5.40 GHz | 5.00 GHz |

| TDP | 45W | 28W |

| Socket | Intel BGA 1744 | AMD Socket FP8 |

| Architecture | Raptor Lake | Zen 5 |

| Codename | Raptor Lake-H | Strix Point |

| Process Node | 10nm | 4nm |

| Foundry | Intel | TSMC |

| Die Size | 257 mm² | 233 mm² |

| L2 Cache | 2 MB (per core) | 1 MB (per core) |

| L3 Cache | 24 MB (shared) | 16 MB |

| Memory Support | DDR4, DDR5 | DDR5, LPDDR5X |

| Memory Bandwidth | Not listed | 89.6 GB/s |

| PCIe | Gen 5, 8 Lanes | Gen 4, 16 Lanes |

| Integrated Graphics | Iris Xe Graphics 96EU | Radeon 880M |

| Release Date | 2023-01-03 | 2024-06-30 |

| Launch MSRP | $697 | None listed |

| Part Number | SRMHU | 100-000001530 |

FAQ

Q: Which processor has better multi-core performance?

A: The Intel Core i9-13905H wins in Cinebench R23 multi-core with 20034 vs 18698, a 7.1% advantage. It also leads in PassMark multithread by 1.1% (29779 vs 29467) and Cinebench R15 multi-core by 0.5% (2857 vs 2842).

Q: Is the AMD chip faster in single-core tests?

A: It depends on the benchmark. AMD wins PassMark single-thread by 3.6% (3841 vs 3703) and Cinebench R15 single-core by 18.8% (303 vs 246), but Intel wins Cinebench R23 single-core by 1.4% (2020 vs 1992).

Q: How do their overall benchmark averages compare?

A: Intel’s average benchmark score is 40313, AMD’s is 40048, a 0.7% difference. Both rank at the 87th percentile against all CPUs.

Q: What are the biggest performance gaps in either direction?

A: Intel’s largest win is in PassMark physics at 20.5% (2054 vs 1704). AMD’s largest win is in Cinebench R15 single-core at 18.8% (303 vs 246), followed by a 15.2% edge in PassMark extended instructions (25113 vs 21297).

Q: Do these processors have the same thread count?

A: Yes, both have 20 threads. However, Intel achieves this with 14 cores while AMD uses 10 cores.

Q: Which chip supports faster PCIe?

A: Intel supports PCIe Gen 5 with 8 CPU lanes, while AMD supports PCIe Gen 4 with 16 CPU lanes.

Where Each One Wins

The Intel Core i9-13905H is the pick for multi-threaded rendering, physics simulation, floating-point math, and data encryption. Its Cinebench R23 multi-core score of 20034 beats AMD’s 18698, and the 20.5% physics advantage (2054 vs 1704) is decisive for simulation workloads. The 16.9% floating-point lead (73434 vs 62802) makes it the better choice for scientific computing, and the 11.8% encryption edge (20464 vs 18297) suits security-conscious users. It also wins prime number finding by 4.3% (122 vs 117), making it slightly better for integer primality tests.

The AMD Ryzen AI 9 365 wins where instruction diversity and efficiency matter. Its 15.2% lead in extended instructions (25113 vs 21297) is a major advantage for code using AVX-512 or similar extensions. String sorting goes to AMD by 5% (39447 vs 37460), which helps with data processing pipelines. PassMark single-thread performance favors AMD by 3.6% (3841 vs 3703), and the 18.8% Cinebench R15 single-core win (303 vs 246) suggests strong legacy single-threaded performance. Data compression and integer math are essentially tied, but AMD edges both by 0.1% (354510 vs 354330 and 101831 vs 101716 respectively). Combined with its 28W TDP versus Intel’s 45W, AMD is the better choice for battery-conscious mobile builds that still want competitive performance.