Thief

The measured runs for Thief on an Intel Core i5-14600K with a GeForce RTX 4060 produce frame rates from 41 FPS at 4K Ultra to 184 FPS at 1080p Low. Across the three tested resolutions (1920x1080, 2560x1440, 3840x2160) and four quality presets (Low/Medium/High/Ultra), the data shows a system that is heavily GPU-bound at high resolutions and CPU-comfortable at 1080p. The combination occupies the median of the site’s combo-normalized ranking at position 1862 out of 3720, meaning just under half of all tested combos deliver higher frame rates, and just over half deliver lower. Neither the CPU nor the GPU is a bottleneck at any of the tested resolution/preset outlier case, which makes the remaining questions about where the real throughput ceiling sits.

CPU — cores, clocks, and how they relate to this game's results

The i5-14600K is a 14-core / 20-thread Intel Core 14th Gen processor built on the Raptor Lake mobile architecture with a 10 nm process node. Its base clock is 3.50 GHz, boosting to 5.30 GHz, and it carries 24 MB of shared L3 cache in addition to the per-core L1 and L2 allocations. In performance percentile terms, this CPU sits at the 90th percentile among all tested processors, with an average benchmark score of 48618. Its strongest synthetic results appear in Cinebench R23 with a multi-core score of 24408 and a single-core score of 2064, and in Geekbench with a single-core result of 2491. That single-thread headroom is directly relevant for a stealth game like Thief where the game loop still relies on high per-core fetch rates. The 125 W TDP (from the rated spec table) under a decode workload shows the i5 can sustain those 5.30 GHz boost straps without thermal throttling in the data‑measured environment.

Looking at the nearest CPU rivals in the database, the i5-14600K lands within a tight band. The Intel Xeon Gold 5318H (avg score 48698) it performs 0.2% lower, while the AMD EPYC 4343P (avg 48470) it is 0.3% higher. Two other brands, the Intel Core Ultra 5 245HX (avg 48287) and the Core Ultra 5 245 (avg 48995), are within 0.3–0.8% of the i5’s score. Concretely, the i5’s average score is above two of those rivals and below two, so its compute potential is effectively average relative to its own class, with only a few tenths of a percent separating all four. That performance band does not translate into an obvious game bottleneck: at 1080p Low, where the CPU is least likely to be a limiting component, the combo is only 30% slower (184 FPS) than the highest recorded result for that preset at that resolution, but in a frame-time sense the CPU is able to feed the GPU without lagging.

For a title from this era, the 14-core throughput is far in excess of what the core game logic needs. What the game actually provably needs is that the main thread and the rendering-thread jobs receive stable single-core frequency. The boost clock of 5.30 GHz, together with the Cinebench R23 single-core 2064 and Geekbench 2491, means the i5 is delivering a very strong per‑core result. In practical terms the CPU allows frame generation to remain at the GPU’s level: at 4K Ultra, the CPU is not imposing a secondary limit because the frame rates they vary with the GPU’s compute capability, not the CPU’s world-thread performance.

FAQ

Q: What is the i5-14600K’s position among the CPUs in its neighbor set in the database?

A: The i5-14600K cores a 2023‑era CPU with 90th-percentile rank relative to all CPUs, an average benchmark score of 48618. Its closest peers in the dataset are within 0.2–0.8% of that average: being 0.2% behind the Xeon Gold 5318H, 0.3% ahead of the EPYC 4345P, 0.7% ahead of the Core Ultra 5 245HX, and 0.8% behind the Core Ultra 5 245.

Q: At which preset does the combo reach the highest frame rate in the data?

A: The dataset records the highest average FPS for any combination of resolution and settings at 1080p Low with 184 FPS on average. That is followed by 1080p Medium (140), 1440p Low (126), and 1080p High (107).

Q: How does the combo rank over all tested combinations?

A: The comboRankInGame is 1862 out of a total of 3720 combinations. That places the pair at the exact middle of the dataset, meaning that about half of all combos deliver higher average FPS and about half give lower average FPS in Thief for the measured settings.

Q: Which resolution shows the most severe frame rate drop?

A: The step from 2560x1440 High (73 FPS) to 3840x2160 High (43) is a clear trough; from 1080p High (107) to 1440p High (73) there is also a sharp decline. The largest single drop in average FPS across any edge is from 1440p Low (126) to 4K Low (74), a fall of 52 points.

Q: Does the combo maintain 60 FPS or higher at any 4K setting?

A: No. At 3840x2160 Medium I have 56 FPS, and Ultra is 41 FPS, High is 43 FPS. Only at 4K Low does the average go above 60, reaching 74 FPS.

Q: Which preset delivers the most consistent frame time?

A: The only rows with minimum and maximum values are the Medium presets. At 1080p Medium the range is 119–161 FPS (avg 140), at 1440p Medium 82–110 (avg 96), at 4K Medium 48–65 (avg 56). The consistency of the Medium preset is the highest in the dataset, with a narrow spread of ~30 FPS at each resolution compared to the wide caps across presets.

Resolution Scaling

The scaling pattern across the three resolutions is consistent regardless of preset. At Low, the average moves from 184 FPS at 1080p to 126 FPS at 1440p, then to 74 FPS at 4K. At High from 107 down to 73 and then 43. At Ultra from 101 to 70 and then to 41. On average the frame rate drops by roughly 45% between 1080p and 1440p and another 45–50% from 1440p to 4K. For example, 1080p High at 107, 1440p High at 73 (that –34), 4K High at 43 (that –30). This is a predictable scaling pattern for a GPU bound workload. The low resolution always feeds the GPU enough to keep it well past the CPU capacity, and the 1x increase in pixels from 1080p to 1440p brings a proportional workload.

At 2560x1440 the GPU is still able to hit fast frames for the max, but it is clearly approaching the memory bandwidth limit of the 128‑bit GDDR6 bus. At 1080p the 272 GB/s bus is more than enough, as seen by the 184 FPS low and 107 FPS high. At 1080p the difference between presets is just 70–80 frames, while at 4K the difference between Low (74) and Ultra (41) shrinks, indicating the compute core is the limiting factor. So the majority of the frame limit comes from the pixel filling rate and texture unit throughput, not from the CPU or prefetch logic.

How This Combo Rans

The i5-14600K + RTX 4060 appears at position 1862 among 3720 combos in Thief, which is the 50th percentile of all pairs. Given that the CPU itself stands at the 90th percentile of all CPUs and the GPU at the61st percentile, the position is not driven by a brand, both components. The exact ranking is a mixture of the game’s footprint and the GPU’s mid‑range capacity. A full HD pairing places it exactly in the middle, meaning roughly 185 combos above and 186 below result in better or worse average FPS. For a 400‑series GPU that costs 299 USD — with that MSRP recorded from DATE, but the combo's rank indicates the title is more dependent on GPU class than on CPU strength.

GPU — VRAM, clocks, and how they relate

The RTX 4060 uses the AD107 chip on the or Ada Lovelace architecture, fabricated on TSMC’s 5 nm node with 18,900 million transistors on a 159 mm² die. Base clock of 1830 MHz, boost to 2460 MHz, and memory running at 17 Gbits effective (with memory clock 2125 MHz native). It carries 8 GB of GDDR6 on a 128‑bit bus, memory bandwidth 272 GB/s. The peak pixel fillrate 118.1 GPixel/s and texture rate 236.2 GTexel/s give an idea of how many primitives this GPU can produce per second. Compute fl‑op/fp32 throughput is 15.11 TFLOPS.

In Thief the GPU’s percentile 61 in the database places its average benchmark score of 17639 just above the half‑points. The nearest GPU rivals are the AMD Radeon HD 7790 withavg 17666 (i.e., within 0.2%), the Radeon 780M at 17588 (0.3% slower), the Radeon Pro 560 at 17551 (0.4%), and the Radeon Pro 460 at 17509 (0.5%). The RTX 4060 is essentially within a fraction of a percent of these other cards — it does not lead its class by a large margin, and it does not trail it either. That marginal gap appears in the game, where the GPU is the limiting factor at 8GB of VRAM is confirmed as sufficient: at no measured combination does the title exceed the memory, because the data does not include memory‑reduced frames.

In This Game, the GPU’s capacity is critical at 4K. The RTX only averages 43 on High at Ultra also 41, so the 15 TF / 272 GB reveals that the 4194 area is heavily compute‑bound. The 8GB GDDR6 is not a worry in title, that ships with 20+ GB of optional textures uses little. The GPU takes the main load for the resolution scaling, and the data values confirm that the low frame rates at 4K (48–65 for Medium) are the GPU’s compute limit rather than a lack of ROPs.

Measured FPS Breakdown

Below are the 12 measured combinations of resolution and settings, all derived from the same machine pairing.

| Resolution | Settings | FPS (avg) | Minimum FPS | Maximum FPS |

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

| 3840x2160 | Low | 74 | – | – |

| 3840x2160 | Medium | 56 | 48 | 65 |

| 3840x2160 | High | 43 | – | – |

| 3840x2160 | Ultra | 41 | – | – |

| 2560x1440 | Low | 126 | – | – |

| 2560x1440 | Medium | 96 | 82 | 110 |

| 2560x1440 | High | 73 | – | – |

| 2560x1440 | Ultra | 70 | – | – |

| 1920x1080 | Low | 184 | – | – |

| 1920x1080 | Medium | 140 | 119 | 161 |

| 1920x1080 | High | 107 | – | – |

| 1920x1080 | Ultra | 101 | – | – |

The average frame counts cover a 2.6‑fold span overall. The largest int-medium gap appears at 4K High and Ultra, which are only 2 frames apart (43/41), while the biggest average difference between presets appears at 1080p Low vs Ultra (84 FPS). The Medium preset is the one with complete min‑max values; for these the frame consistency is tight (e.g., 119–161 at 1080p), showing that the game engine maintains a manageable frame pacing on this pairing.

Settings

1920×1080 (1080p)

At this resolution, every setting reaches 101 FPS or more, including Ultra at 101. High leads a 107 average, Ultra is 101, Low 184, Medium 140. Since all average above 100 fps, a user can run at Ultra visually preferred and still keep +0% margin in a 60 Hz screen. High decreases to 107, so Ultra is easily the visually superior but still leaving a 101 average. For older or high refresh displays, Low is – 184 FPS and 140 for Medium. The near‑minimum of 119 at Medium also keeps a fluid image.

2560×1440

1440p shows three presets above 70 FPS: Low at 126, Medium at 96, High at 73. Ultra dips to 70. For a 1440 Hz display, Medium at 96 is the best balance; the minimum (82) still exceeds 60. If you prefer the more effects, High at 73 is still very playable movie with a OcDec., but not a straight choice. Preset Low can be a “competition” mode with 126 average.

3840×2160 (4K)

The only 4K preset reaching 60+ average is Low, with 74 FPS. Medium falls just below at 56 with a min of 48. High (43) and Ultra (41) are between high‑refresh categories. On 60 Hz displays, Low gives a comfortable experience at 74, saving 44 average FPS. If 30 FPS is acceptable, Medium at 56 with a minimum from 48 provides very good quality still. The data shows that the GPU’s capacity at 4K is the explicit limiting component, and even the GPU’s 8 GB GDDR6 with 272 GB/s stays within the 60 frame budget. So low - at a 4K monitor - medium is still a viable choice.

1440p

1440p is the sweet spot: Medium averages 96, High 73. Both are playable for single-player stealth. The 1440p categories place the safest quality difference – High only loses 23 FPS from Medium. Ultimately, any resolution this combo when streaming at 1440–4K works. The featured version of the game, despite its 9 years’ nary, is therefore