Instant compatibility check with FPS benchmarks and optimization tips
Thief
Performance
1080p (Full HD)
1440p (2K / QHD)
4K (Ultra HD)
Requirements Check
Recommendations
Your hardware doesn't meet minimum requirements for this game.
Your GPU supports ray tracing.
Average FPS: resolution vs quality settings
| Ultra | High | Medium | Low | |
|---|---|---|---|---|
| 4K Ultra HD | 74 | 75 | 83 | 92 |
| 1440p QHD | 89 | 90 | 99 | 110 |
| 1080p Full HD | 100 | 102 | 113 | 125 |
Cell color: green is 120+ FPS, teal is 60+, amber is 30+, red is below 30. Cards marked Estimated above have no measured row and are scaled from your GPU benchmark percentile.
Performance Analysis: Thief on Your System
Similar Performance Alternatives
The recorded benchmark data places this Intel Core i5-10400F configuration near a distinct cluster of processors. The database shows the i5-10400F with an average benchmark score of 14,185, which puts it at the 68th percentile among all CPUs. The nearest rivals in the database are separated by less than one percent in either direction. The Intel Xeon 6756E scores 14,163, a delta of 0.2 percent, meaning it is effectively tied with the i5-10400F in aggregate performance. The Intel Core 7 160UL scores 14,232, which is 0.3 percent higher, again a negligible difference. On the AMD side, the EPYC 7552 scores 14,115 at 0.5 percent lower, and the Ryzen 3 7320C scores 14,277 at 0.6 percent higher. For practical purposes, any of these processors would deliver nearly indistinguishable frame pacing in this game, as the aggregate scores fall within a 1.2 percent band.
The GPU pairing is more interesting. The NVIDIA GeForce RTX 5090 D has an average benchmark score of 77,712 and sits at the 92nd percentile among all GPUs. Its nearest rivals in the database are not the expected high-end cards, but rather a set of mobile and workstation parts. The AMD Radeon RX 6650M XT scores 76,904, which is 1.1 percent lower. The AMD Radeon RX 6850M XT scores 78,940, which is 1.6 percent higher. The NVIDIA Tesla P100 PCIe 12 GB and 16 GB variants score 79,396 and 79,605 respectively, placing them 2.1 and 2.4 percent above the RTX 5090 D. This is a counterintuitive result: the desktop RTX 5090 D, with its 32 GB of GDDR7 memory and 21,760 shading units, benchmarks near these parts in the database's aggregate metric. That said, the game-specific measurements below show that this GPU is more than capable for Thief, so the aggregate score is not the whole story.
For the game itself, the data shows that this CPU and GPU combo is far from the top of the database. The combo rank is 1,684 out of 3,720 tested combinations, which places it in the 55th percentile of all tested pairs. This is a notable mismatch: the GPU is in the 92nd percentile, the CPU is in the 68th percentile, yet the combo lands lower than both individual percentiles would suggest. The measured frame rates, however, tell a different story, as the game is not particularly demanding at the settings tested.
CPU and GPU Roles
The measured FPS data reveals which component matters more for Thief at each resolution. At 1920x1080, the resolution most sensitive to CPU performance, the frame rates are the highest recorded. At Low settings, the average FPS is 125. At Medium, it drops to 113. At High, it is 102. At Ultra, it is 100. The spread from Low to Ultra is 25 FPS, which is the narrowest of the three resolutions. This suggests that at 1080p, the GPU is not the limiting factor, and the CPU is able to feed the GPU adequately at all settings. The i5-10400F, with its 6 cores and 12 threads, appears sufficient for this title at 1080p.
At 2560x1440, the frame rates step down. Low averages 110 FPS, Medium averages 99 FPS, High averages 90 FPS, and Ultra averages 89 FPS. The drop from 1080p Low to 1440p Low is 15 FPS, a 12 percent reduction. The drop from 1080p Ultra to 1440p Ultra is 11 FPS, an 11 percent reduction. This scaling indicates that the GPU is beginning to take on more load, but the CPU is still keeping pace. The delta between Low and Ultra at 1440p is 21 FPS, slightly narrower than at 1080p.
At 3840x2160, the GPU becomes the dominant factor. Low averages 92 FPS, Medium averages 83 FPS, High averages 75 FPS, and Ultra averages 74 FPS. The drop from 1440p Low to 4K Low is 18 FPS, a 16 percent reduction. The drop from 1440p Ultra to 4K Ultra is 15 FPS, a 17 percent reduction. The spread from Low to Ultra at 4K is 18 FPS, the narrowest of all resolutions. This pattern is consistent with a GPU-bound scenario: as resolution increases, the GPU's fill rate and memory bandwidth become the limiting factors, and the CPU's role diminishes. The RTX 5090 D, with its 1.79 TB/s memory bandwidth and 423.6 GPixel/s pixel rate, is clearly the component doing the heavy lifting at 4K.
The data also shows that the game's presets have a modest impact on performance. At every resolution, the difference between Low and Ultra is less than 25 percent of the Low score. This suggests that Thief's Ultra preset is not dramatically more demanding than its Low preset, at least in the scenes measured. The min and max FPS values, available only for Medium settings, reinforce this. At 1080p Medium, the range is 96 to 130 FPS. At 1440p Medium, it is 85 to 114 FPS. At 4K Medium, it is 71 to 96 FPS. These ranges are tight, indicating stable frame delivery without severe dips.
FAQ
Q: Can this PC run Thief at 4K with the Ultra preset?
A: Yes. The measured average FPS at 3840x2160 with the Ultra preset is 74 FPS, which is above the 60 FPS playable threshold. The verdict data confirms that Ultra is the best playable setting at this resolution.
Q: What is the highest average frame rate recorded for this combo?
A: The highest average FPS is 125, recorded at 1920x1080 with the Low preset. This is the fastest measurement in the entire dataset.
Q: How does the CPU compare to its nearest rivals?
A: The Intel Core i5-10400F has an average benchmark score of 14,185. Its nearest rival, the Intel Xeon 6756E, scores 14,163, which is 0.2 percent lower. The AMD EPYC 7552 scores 14,115, which is 0.5 percent lower. The performance difference is under one percent in all cases.
Q: Is the GPU or CPU more important at 1080p?
A: At 1920x1080, the CPU plays a larger role. The frame rates are highest at this resolution, and the drop from Low to Ultra is only 25 FPS, suggesting the GPU is not saturated. The CPU's 6-core, 12-thread configuration is sufficient to drive the GPU at all presets.
Q: What is the minimum frame rate recorded at 1440p?
A: The only min FPS values recorded are for the Medium preset. At 2560x1440 Medium, the min FPS is 85, the max is 114, and the average is 99. No min or max values are recorded for other presets at this resolution.
Q: How does this combo rank among all tested combinations?
A: The combo rank is 1,684 out of 3,720 tested combinations. This places it in the middle of the database, although the game-specific performance is strong at all resolutions.
The Verdict
The data is unambiguous: this PC can run Thief at every resolution and preset tested, and it clears the 60 FPS playable threshold by a wide margin. At 1920x1080, the lowest average FPS is 100, recorded at the Ultra preset. At 2560x1440, the lowest average FPS is 89, again at Ultra. At 3840x2160, the lowest average FPS is 74, also at Ultra. The verdict by resolution confirms that all four settings, Low, Medium, High, and Ultra, are playable at all three resolutions. The best settings per resolution are Ultra at each one, with average FPS of 100 at 1080p, 89 at 1440p, and 74 at 4K.
The recorded data shows no scenario where this combo dips below 60 FPS. Even at the most demanding preset and resolution, the average FPS is 74, which is 23 percent above the playable threshold. The game is from 2014, and the hardware is from 2020 and 2025, so this result is unsurprising, but the data confirms it. For a 60 FPS target, this system has ample headroom at every setting. For a 144 Hz monitor at 1080p, the Low preset at 125 FPS would be the only way to approach that refresh rate, but the Ultra preset at 100 FPS is still smooth.
How This Combo Ranks
The combo rank in the database is 1,684 out of 3,720, which places it in the 55th percentile of all tested combinations. This rank is lower than the individual component percentiles would suggest. The CPU is at the 68th percentile, and the GPU is at the 92nd percentile. A naive expectation would place the combo higher, but the database's rank is based on the aggregate performance across all tested games, not just Thief. In this specific title, the measured FPS is high, but the rank reflects a broader set of workloads.
The nearest rival CPUs, as listed in the database, have average scores within 0.6 percent of the i5-10400F. The nearest rival GPUs, the RX 6650M XT and RX 6850M XT, are within 1.6 percent of the RTX 5090 D. If a user had any of these alternate parts, the frame rates in Thief would be nearly identical. The data shows that the CPU is not a bottleneck in this game, and the GPU is more than sufficient, so swapping to a rival part within that performance band would not change the outcome.
The rank also indicates that there are many faster combinations in the database, but for this specific game, the rank does not reflect the playability. A combo ranked in the middle of the pack can still deliver 74 FPS at 4K Ultra, which is well above the threshold. The rank is a global metric, not a game-specific one.
Best Settings per Resolution
The most demanding preset that stays at or above 60 FPS at each resolution is Ultra, across the board. At 1920x1080, the Ultra preset averages 100 FPS. This is the highest quality setting available, and it delivers a smooth experience. At 2560x1440, the Ultra preset averages 89 FPS. Again, this is the maximum quality setting, and it clears the threshold by nearly 30 FPS. At 3840x2160, the Ultra preset averages 74 FPS. This is the most demanding combination of resolution and settings in the dataset, and it still remains above 60 FPS.
The verdict data confirms that all four presets are playable at all three resolutions, so there is no need to compromise on quality. The only reason to choose a lower preset would be to reach higher refresh rates. For example, at 1080p, the Low preset averages 125 FPS, which would suit a 120 Hz display. At 1440p, the Low preset averages 110 FPS, which would also suit a 120 Hz display. At 4K, the Low preset averages 92 FPS, which is below 100 but still high. The database does not record any preset that drops below 60 FPS, so the recommendation is straightforward: use Ultra at any resolution for the best visual quality.
Measured FPS Breakdown
The measured data is divided by resolution and settings, with average FPS recorded for every combination. Min and max FPS are only recorded for the Medium preset at each resolution.
At 1920x1080, the results are as follows. Low averages 125 FPS. Medium averages 113 FPS, with a min of 96 and a max of 130. High averages 102 FPS. Ultra averages 100 FPS. The total spread from Low to Ultra is 25 FPS. The Medium preset's min of 96 FPS is the lowest recorded value at this resolution, but it remains well above the 60 FPS threshold.
At 2560x1440, the results step down. Low averages 110 FPS. Medium averages 99 FPS, with a min of 85 and a max of 114. High averages 90 FPS. Ultra averages 89 FPS. The spread from Low to Ultra is 21 FPS. The Medium preset's min of 85 FPS is the lowest recorded value at this resolution, still comfortably above 60 FPS.
At 3840x2160, the results are the lowest of the three. Low averages 92 FPS. Medium averages 83 FPS, with a min of 71 and a max of 96. High averages 75 FPS. Ultra averages 74 FPS. The spread from Low to Ultra is 18 FPS. The Medium preset's min of 71 FPS is the lowest recorded value across the entire dataset, but it is still 11 FPS above the playable threshold.
The data shows a consistent scaling pattern. Moving from 1080p to 1440p reduces average FPS by roughly 10 to 12 percent at each preset. Moving from 1440p to 4K reduces average FPS by roughly 16 to 17 percent. The GPU's memory bandwidth and pixel rate, as recorded in the specification, are the likely drivers of this scaling. The CPU's performance, measured at 4,748 in the 3DMark 16-thread test and 688 in the single-thread test, is sufficient to avoid bottlenecking at any resolution tested.