Intel Core i9-12900F vs Intel Core Ultra 5 235 Comparison
Intel Core i9-12900F
Core Ultra 5 235
PERFORMANCE BENCHMARKS
Analysis: Intel Core i9-12900F vs Intel Core Ultra 5 235
FAQ
Q: Which processor has more cores and threads?
A: The Intel Core i9-12900F has 16 cores and 24 threads, while the Intel Core Ultra 5 235 has 14 cores and 14 threads. The i9-12900F offers 2 additional cores and 10 additional threads.
Q: What are the benchmark score distributions between these two CPUs?
A: The i9-12900F wins 8 of the 17 head-to-head benchmarks, while the Core Ultra 5 235 wins 9. The average benchmark score for the i9-12900F is 47,176, and for the Core Ultra 5 235 it is 46,062, a gap of roughly 2.4%.
Q: Which processor has the higher boost clock?
A: The Intel Core i9-12900F has a boost clock of 5.10 GHz, which is 0.10 GHz higher than the Core Ultra 5 235's boost clock of 5.00 GHz. The Core Ultra 5 235 has a higher base clock at 3.40 GHz compared to 2.40 GHz for the i9-12900F.
Q: Do both processors support the same memory types?
A: No. The i9-12900F supports both DDR4 and DDR5 memory, while the Core Ultra 5 235 supports only DDR5. The Core Ultra 5 235 has a higher memory bandwidth at 102.4 GB/s, compared to 76.8 GB/s for the i9-12900F.
Q: Which processor includes integrated graphics?
A: Only the Intel Core Ultra 5 235 includes integrated graphics, specifically Arc Xe-LPG Graphics 24EU. The i9-12900F has no integrated graphics listed in the data.
Q: What are the process nodes for these two CPUs?
A: The i9-12900F is built on Intel's 10 nm process, while the Core Ultra 5 235 is built on TSMC's 3 nm process. The Core Ultra 5 235 also has a smaller die at 243 mm², though it packs 17,800 million transistors.
Architecture Differences
The Intel Core i9-12900F belongs to the Alder Lake generation, specifically Alder Lake-S, and uses the Intel Socket 1700. It is built on Intel's 10 nm process node with a die size of 215 mm². The Core Ultra 5 235 belongs to the Arrow Lake generation, specifically Arrow Lake-S, and uses Intel Socket 1851. It is built on TSMC's 3 nm process with a die size of 243 mm² and an explicit transistor count of 17,800 million.
The cache hierarchy differs substantially. The i9-12900F has 80 KB of L1 cache per core, 1.25 MB of L2 cache per core, and 30 MB of shared L3 cache. The Core Ultra 5 235 has 192 KB of L1 cache per core, 3 MB of L2 cache per core, and 24 MB of shared L3 cache. Although the Core Ultra 5 235 has more L1 and L2 per core, its total L3 pool is 6 MB smaller.
The core topology is also different. The i9-12900F provides 16 cores with 24 threads, meaning it supports hyper-threading. The Core Ultra 5 235 provides 14 cores with 14 threads, meaning each core maps to exactly one thread, reflecting a simpler threading model.
Memory support separates the two significantly. The i9-12900F supports both DDR4 and DDR5 in a dual-channel configuration, with a memory bandwidth of 76.8 GB/s. The Core Ultra 5 235 supports only DDR5 in a dual-channel configuration, but achieves a higher memory bandwidth of 102.4 GB/s. The i9-12900F supports ECC memory, while the Core Ultra 5 235 does not.
PCIe connectivity also differs. The i9-12900F provides Gen 5 with 16 lanes (CPU only), while the Core Ultra 5 235 provides Gen 5 with 20 lanes (CPU only). The Core Ultra 5 235 includes an integrated GPU, the Arc Xe-LPG Graphics 24EU, whereas the i9-12900F has no integrated graphics listed.
The i9-12900F has an unlocked multiplier, while the Core Ultra 5 235 is locked. The i9-12900F launched in January 2022, and the Core Ultra 5 235 launched in January 2025, a three-year gap in release timing.
Head-to-Head Benchmarks
The most dramatic differences appear in the Cinebench suite, where the i9-12900F dominates across the board. In Cinebench R15 multi-core, the i9-12900F scores 3,064 against 1,488 for the Core Ultra 5 235, a 105.9% advantage. The single-core R15 result follows the same pattern: 432 against 210, a 105.7% delta. This pattern repeats exactly in Cinebench R20 and R23, with the i9-12900F leading by 105.9% in both multi-core and single-core tests. Specifically, R20 multi-core shows 12,770 versus 6,202, and R23 multi-core shows 30,405 versus 14,769.
The PassMark suite tells a more mixed story. The i9-12900F wins in integer math with a score of 129,504 against 87,948, a 47.3% margin. It also wins in data compression, scoring 451,402 against 390,711, a 15.5% lead. However, the Core Ultra 5 235 takes several notable victories. In find prime numbers, it scores 371 against 127, a 65.8% advantage for the Core Ultra. In floating point math, it scores 117,951 against 96,452, an 18.2% win. The Core Ultra 5 235 also leads in physics (2,570 versus 1,842, a 28.3% margin), data encryption (29,293 versus 25,251, a 13.8% margin), and extended instructions (32,752 versus 28,265, a 13.7% margin).
The single-thread PassMark result favors the Core Ultra 5 235, with 4,516 versus 4,017, an 11% lead. The multithread PassMark score also goes to the Core Ultra 5 235, 37,816 versus 35,912, a 5% margin. Random string sorting is nearly tied, with the Core Ultra 5 235 ahead by only 1% (48,980 versus 48,477).
The overall benchmark count is close: the i9-12900F wins 8 tests, and the Core Ultra 5 235 wins 9 tests. The i9-12900F's wins are concentrated in rendering workloads and integer-heavy tasks, while the Core Ultra 5 235 wins in encryption, floating point, and several single-threaded workloads.
Specification Differences
The two processors differ across most core specifications. The i9-12900F has 16 cores and 24 threads, while the Core Ultra 5 235 has 14 cores and 14 threads. Base clocks are 2.40 GHz for the i9-12900F and 3.40 GHz for the Core Ultra 5 235. Boost clocks are 5.10 GHz and 5.00 GHz, respectively.
Both have a TDP of 65, but they use different sockets: Intel Socket 1700 for the i9-12900F and Intel Socket 1851 for the Core Ultra 5 235. The process nodes differ (10 nm for the i9-12900F, 3 nm for the Core Ultra 5 235), as do the foundries (Intel for the i9-12900F, TSMC for the Core Ultra 5 235). The die sizes are 215 mm² and 243 mm², respectively.
Cache configurations differ in every level: L1 is 80 KB per core for the i9-12900F versus 192 KB per core for the Core Ultra 5 235; L2 is 1.25 MB per core versus 3 MB per core; L3 is 30 MB shared versus 24 MB shared.
Memory support diverges: the i9-12900F supports DDR4 and DDR5, while the Core Ultra 5 235 supports only DDR5. Memory bandwidth is 76.8 GB/s for the i9-12900F and 102.4 GB/s for the Core Ultra 5 235. ECC memory is supported only on the i9-12900F.
PCIe lanes differ: 16 lanes for the i9-12900F and 20 lanes for the Core Ultra 5 235, both Gen 5. The i9-12900F has no integrated graphics, while the Core Ultra 5 235 has Arc Xe-LPG Graphics 24EU. The i9-12900F has an unlocked multiplier; the Core Ultra 5 235 is locked.
The Verdict
The data points to two distinct usage profiles. The Intel Core i9-12900F is the clear choice for multi-core rendering workloads. Its Cinebench R23 multi-core score of 30,405 is more than double the Core Ultra 5 235's 14,769, a 105.9% advantage. Anyone running Cinebench-style workloads, which typically represent 3D rendering and content creation, should favor the i9-12900F without hesitation.
The Intel Core Ultra 5 235, however, wins in several single-threaded and specialized workloads. Its PassMark single-thread score of 4,516 is 11% higher than the i9-12900F's 4,017. It also wins in physics by 28.3%, in floating point math by 18.2%, and in find prime numbers by 65.8%. These results indicate the Core Ultra 5 235 has superior per-core efficiency in certain mathematical and scientific tasks, despite its lower core count.
The average benchmark scores are close, with the i9-12900F at 47,176 and the Core Ultra 5 235 at 46,062, a difference of only about 2.4%. Both processors sit at the 89th percentile among all CPUs, meaning they are statistically equivalent in overall standing. The i9-12900F's nearest rival, the Intel Core i7-13700KF, is 0.3% ahead, while the Core Ultra 5 235's nearest rival, the AMD Ryzen AI 9 HX 375, is 0.1% behind.
The decision comes down to workload. For rendering, the i9-12900F is the decisive winner. For mixed workloads that include encryption, floating point math, and single-threaded applications, the Core Ultra 5 235 shows measurable advantages. The i9-12900F also supports ECC memory and both DDR4 and DDR5, which may matter for specific system configurations. The Core Ultra 5 235 offers integrated graphics, a newer 3 nm process, and higher memory bandwidth.
Where Each One Wins
The i9-12900F wins in all Cinebench tests, making it the stronger processor for rendering workflows. It also wins in PassMark integer math (47.3% ahead) and data compression (15.5% ahead). These are tasks that benefit from a high core count and hyper-threading. The i9-12900F's 24 threads give it a structural advantage in heavily parallel workloads, and the data confirms this with 105.9% leads across all Cinebench versions.
The Core Ultra 5 235 wins in PassMark find prime numbers (65.8% ahead), physics (28.3% ahead), floating point math (18.2% ahead), data encryption (13.8% ahead), and extended instructions (13.7% ahead). It also takes the single-thread PassMark test (11% ahead) and the multithread PassMark test (5% ahead). The Core Ultra 5 235's wins are in areas that respond to architectural efficiency, newer process technology, and higher per-core performance rather than raw thread count.
The Core Ultra 5 235 also wins in random string sorting, though by a narrow 1% margin. The i9-12900F does not win any single-threaded benchmark in the recorded data. The Core Ultra 5 235's higher base clock of 3.40 GHz, compared to 2.40 GHz for the i9-12900F, likely contributes to its single-thread and latency-sensitive wins.
For users who prioritize multi-threaded rendering, the i9-12900F is the clear pick. For users who prioritize encryption, floating point math, or single-threaded responsiveness, the Core Ultra 5 235 has the edge. The Core Ultra 5 235 also includes integrated graphics, which could eliminate the need for a discrete GPU in basic display configurations. The i9-12900F, with its ECC memory support and DDR4 compatibility, may be preferable for certain server or workstation builds, though its lack of integrated graphics is a potential drawback.