Valorant

Valorant paired with the Intel Core Ultra 5 125H and NVIDIA GeForce RTX 5090 is a pairing that the measured data shows to be heavily GPU-underutilized in the classic sense — and that is exactly what you want for a competitive FPS. Across every tested resolution and preset, the combination delivers frame rates in the high hundreds per second, from a floor of 370 avg FPS at 4K Ultra to a peak of 509 avg FPS at 1080p Low. The RTX 5090, sitting in the 92nd percentile of all tested GPUs, never breaks a sweat rendering a game released in 2020, while the Meteor Lake mobile CPU contributes the single-thread speed Valorant's engine actually depends on. The verdict: this combo clears every competitive-play threshold by an enormous margin, and settings choice is about preference, not necessity.

CPU Role

The Intel Core Ultra 5 125H is a 14-core, 18-thread Meteor Lake mobile processor with a 3.60 GHz base clock and a 4.50 GHz boost clock, built on Intel's 7 nm process. For Valorant, the numbers that matter most are the single-core results: a Cinebench R23 single-core score of 1647.5 and a PassMark single-thread score of 3399. Valorant is a competitive FPS that leans on fast individual threads to drive its high frame rates, and a 4.50 GHz boost on a modern architecture is squarely in the territory where frame delivery becomes smooth rather than stutter-prone.

The chip's multi-core credentials — a Cinebench R23 multi-core score of 12804.5, a Cinebench R20 multi score of 7161, and a PassMark multithread score of 21100 — matter less for Valorant itself, but they ensure background tasks, voice chat, and streaming do not steal frames. In aggregate benchmark standing, the CPU sits in the 79th percentile versus all tested CPUs with an average benchmark score of 27507. Its nearest rivals tell the story of a chip locked in a tight performance band: the AMD Ryzen 7 5800 scores 27535 on average (0.1% ahead), the Intel Core i5-12600K and AMD Ryzen 7 5700X both average 27578 (0.3% ahead), and the AMD Ryzen 3 PRO 5355G averages 27382 (0.5% behind). In other words, the Ultra 5 125H trades blows within a fraction of a percent of desktop mainstays — a strong showing for a 28 W mobile part.

How does this translate to the game? With an RTX 5090 pushing 400-plus frames at all times, the CPU becomes the practical ceiling. The compressed spread between 1080p Low (509 avg FPS) and 4K Ultra (370 avg FPS) indicates the frame rate is far less resolution-sensitive than a GPU-bound title would be, which is consistent with the processor and game engine setting the pace. The 18 MB of shared L3 cache and dual-channel DDR5 memory support with 89.6 GB/s of bandwidth feed the engine adequately, and the 28 W TDP means this is a thermally modest chip driving a flagship GPU — an unusual but effective asymmetry.

Settings Recommendations

The measured rows make one thing clear: there is no bad preset here. The data shows remarkably flat scaling across settings, which means the recommendation can be driven entirely by frame consistency rather than raw output.

At 1920x1080, Low delivers 509 avg FPS, Medium 434 (with a 369 minimum and 500 maximum), High 427, and Ultra 420. At 2560x1440, the range runs from 483 avg FPS on Low down to 399 on Ultra, with Medium at 412 (350 minimum, 474 maximum). At 3840x2160, Low hits 448 avg FPS while Ultra still manages 370, with Medium at 382 (325 minimum, 440 maximum).

For competitive play, Medium is the standout pick. It preserves the majority of the Low-preset frame rate while offering better visual clarity than the stripped-down option, and it is the only preset with recorded min/max data in this dataset — the 350 minimum at 1440p Medium and 369 minimum at 1080p Medium both sit comfortably above any threshold where frame drops would affect aim tracking. The gap between the 1080p Low high of 509 and the Ultra result of 420 is less than 90 frames across the entire settings ladder; that is a narrow band for a game where every preset already runs far beyond what high-refresh displays can show.

For visual fidelity, Ultra is entirely viable: 420 avg FPS at 1080p and 399 at 1440p mean even maximum settings leave enormous headroom. The Ultra-versus-Low trade at 1440p is 84 frames — real, but irrelevant to playability. The data supports running Ultra at 1440p or 4K for casual queues and dropping to Medium or Low for ranked, with the confidence that no measured row in this dataset falls below a 325 average.

FAQ

Q: Can this combo run Valorant at 4K with high frame rates?

A: Yes, and decisively. At 3840x2160, the measured averages are 448 FPS on Low, 382 on Medium, 376 on High, and 370 on Ultra. Even the most demanding 4K configuration stays well above the frame rates most monitors can display.

Q: Is the Core Ultra 5 125H a bottleneck for Valorant with an RTX 5090?

A: Not in any way that matters. The CPU's 1647.5 Cinebench R23 single-core score and 3399 PassMark single-thread score place it within 0.3% of chips like the Core i5-12600K and Ryzen 7 5700X in aggregate performance, and the measured frame rates never dip below 370 avg FPS at any setting. If a ceiling exists, it is far above competitive requirements.

Q: How much FPS do you gain by dropping from Ultra to Low at 1080p?

A: The measured difference is 89 frames: 420 avg FPS on Ultra versus 509 on Low. That is the full settings range at the most CPU-sensitive resolution, and it confirms the game is not GPU-limited with this hardware.

Q: Does raising the resolution hurt frame rate significantly?

A: No. At Ultra settings, going from 1080p to 4K costs only 50 avg FPS (420 to 370). That shallow resolution scaling is characteristic of a CPU/engine-limited scenario rather than a GPU-limited one, despite the RTX 5090's enormous headroom.

Q: What is the worst measured frame rate in this dataset?

A: The lowest recorded average is 370 FPS at 3840x2160 Ultra. The lowest recorded instantaneous value is the 325 minimum at 4K Medium.

Q: How does the RTX 5090 compare to rival GPUs in the database?

A: Its average benchmark score of 79842 places it 0.3% ahead of the NVIDIA Tesla P100 PCIe 16 GB (79605), 0.6% ahead of the Tesla P100 PCIe 12 GB (79396), 1.1% ahead of the AMD Radeon RX 6850M XT (78940), and 1.4% behind the AMD Radeon Pro Vega 64X (80959) in aggregate standing — while sitting in the 92nd percentile of all GPUs.

How This Combo Ranks

Within this game's database, the Core Ultra 5 125H plus RTX 5090 pairing ranks 1422 out of 3708 tested combinations. That places it in the top half of all tested pairings — better than the majority of combos — but not at the top, which reflects how Valorant rewards raw single-thread CPU speed over GPU horsepower. A game that outputs 370 avg FPS at 4K Ultra does not differentiate between GPUs the way modern AAA titles do; a stronger CPU in the pairing would climb the leaderboard faster than a stronger GPU.

Interpreted practically: the rank data and the FPS data tell the same story from two angles. The RTX 5090 is overkill for Valorant specifically, and the leaderboard position of this combo is capped more by the mobile-class processor than by the flagship graphics card. Players whose priority is Valorant alone could pair the same GPU class tier down and lose nothing measurable; players who also play demanding titles get a combo that ranks solidly while obliterating this game's requirements.

GPU Role

The NVIDIA GeForce RTX 5090 is a Blackwell 2.0 flagship built on TSMC's 5 nm process, with 92,200 million transistors packed into a 750 mm² die at a density of 122.9M per mm². It carries 21760 shading units, 680 texture mapping units, 176 render outputs, 170 RT cores, and 680 tensor cores, with a base clock of 2017 MHz and a boost clock of 2407 MHz. Memory is 32 GB of GDDR7 on a 512-bit bus running at 1750 MHz (28 Gbps effective), delivering 1.79 TB/s of bandwidth. The raw compute figures — 104.8 TFLOPS FP32, 423.6 GPixel/s pixel rate, and 1,636.8 GTexel/s texture rate — describe a card engineered for far heavier workloads than a 2020 competitive shooter.

In Valorant, the result of that overkill is visible in the resolution scaling. The step from 1080p to 4K is a quadrupling of pixel count, yet the measured Ultra frame rate only drops from 420 to 370. When a GPU has this much headroom, the game's frame rate becomes limited by the CPU and engine instead, which is precisely what the flat scaling across presets and resolutions demonstrates. The card's benchmark standing reinforces this: a 3DMark Steel Nomad DX12 score of 18355, a Geekbench Vulkan score of 376728, a Geekbench OpenCL score of 334370, and a PassMark G3D score of 39650 place it in the 92nd percentile of all tested GPUs with an average score of 79842.

None of this GPU's strengths are challenged by Valorant. The 32 GB frame buffer is untouched by the game's modest asset pool, and the API support — DirectX 12 Ultimate, Vulkan 1.4, OpenGL 4.6 — comfortably exceeds what the title requires. The card draws 575 W TDP through a single 16-pin connector with a 950 W suggested PSU, and fits as a dual-slot card measuring 304 mm long — practical considerations for the platform builder, though none of it is exercised meaningfully by this particular game.

Measured FPS Breakdown

At 1920x1080, the combo posts its highest numbers: 509 avg FPS on Low, 434 on Medium with a 369 minimum and 500 maximum, 427 on High, and 420 on Ultra. The Medium row is the only configuration in this dataset with recorded minimum and maximum values, and its 131-frame min-to-max spread at 1080p indicates stable frame delivery rather than erratic spikes.

At 2560x1440, the averages are 483 (Low), 412 (Medium, with a 350 minimum and 474 maximum), 405 (High), and 399 (Ultra). The Low-to-Ultra span here is 84 frames — under 18% of the peak figure — confirming that the settings ladder barely moves the needle. The Medium minimum of 350 sits within 100 frames of its 474 maximum, a tight band for a live multiplayer environment.

At 3840x2160, the results are 448 avg FPS on Low, 382 on Medium (325 minimum, 440 maximum), 376 on High, and 370 on Ultra. Even at maximum resolution and maximum settings, the average stays at 370 — a figure that exceeds what nearly any display can render in full, and one that dwarf typical competitive-play requirements.

Read across the grid, three patterns hold. First, Low-to-Ultra differences are small everywhere: 89 frames at 1080p, 84 at 1440p, 78 at 4K. Second, resolution differences are equally shallow: Ultra loses just 21 frames from 1080p to 1440p and another 29 from 1440p to 4K. Third, every single measured average sits above 370. This is a pairing where the benchmark question is not "is it fast enough" but "how far past enough can it go," and the answer at every tested configuration is: far past.