Final Fantasy 16 on PC is a GPU-bound game, so lowering the graphics preset buys you little and upscaling buys you a lot. PC Optimized Settings found it “fully GPU bound even at 1080p with upscaling enabled” on a Ryzen 9 7950X with an RTX 4090. In a separate test, DSOGaming saw only a 14–19% gain going from Max to Low at native 1080p, while DLSS added 45–85% at 4K in PC Optimized Settings’ run.
That changes the order you tune things in. Resolution and upscaling first, a handful of cheap toggles second, the rest of the menu last. I haven’t benchmarked the game on my own rigs, so every number below is another tester’s, with the hardware and date stated. The general method behind this ordering is in our guide to optimizing PC game settings for better FPS.
Version note: checked against Steam’s version history through update 1.03 (4 March 2025), the latest PC version update on the news feed as of 5 October 2026. Most benchmarks were run in September–October 2024, and update 1.02 changed Shadow Quality and Graphical Fidelity afterwards, so treat the per-setting numbers for those two as pre-patch.
Quick Start: Eight Changes in Order
- Let the shaders compile on first launch. Notebookcheck reports shader compilation takes “several minutes” at first start even on powerful systems. Don’t judge performance until it finishes.
- Set the upscaler to Quality. DLSS on GeForce RTX cards, FSR 3 on any card, XeSS on Intel Arc. This is the largest single gain in the menu.
- Set Graphical Fidelity and Shadow Quality to Medium. That is the pairing Digital Foundry’s optimized preset used.
- Keep Terrain Quality and Water Quality on High. Their measured cost is small.
- Turn Variable Rate Shading off. Digital Foundry did, and PC Optimized Settings found its impact on quality and performance “indetectable.”
- Leave NPC Density, Ambient Occlusion and Bloom alone. None of them moved the frame rate meaningfully in testing.
- Need about 10% more? Lower Screen Space Reflections, then Clutter Density. Those two had the largest measured single-setting gains.
- Add Frame Generation last, and only once the game already runs acceptably without it.
Is Final Fantasy 16 CPU-Bound? No, and Five Sources Agree
No test I found shows Final Fantasy 16 limited by the processor on a reasonable CPU. The table lists what each source measured.
| Source | Test rig | What it found |
|---|---|---|
| PC Optimized Settings | Ryzen 9 7950X, RTX 4090 | “Fully GPU bound even at 1080p”; GPU-Busy deviation of only 1–2% across all benchmarks |
| GGRecon | Core i9-10900K, RTX 3080, 3440×1440 | “Bottle-necked by your graphics card”; CPU around 60% in the most intensive scenes |
| DSOGaming | Ryzen 9 7950X3D, several GPUs | “Does not need a high-end CPU, it certainly does require a powerful GPU” |
| GameGPU | Ryzen 3 3100 up to Ryzen 9 7950X and Core i9-14900K, 1080p, max settings | Lists the Ryzen 3 3100 and Core i3-10100 as sufficient for both its “acceptable” and “comfortable” FPS tiers |
| Digital Foundry (via Multiplayer.it) | Ryzen 3600 + RTX 2070 Super; RTX 4090 + Core i7-13700K | Called “one of the heaviest GPU-side games of this generation” |
The pattern across the table: every test that tried to expose a processor limit ended up describing a graphics-card limit instead. The game also spreads its work well. GameGPU found it can load up to 16 threads and uses about 12 of them efficiently, and DSOGaming called the CPU scaling “exceptional.”
That includes the setting that looks most like a CPU cost. NPC Density controls crowds, which is exactly what you would expect to load a processor, yet PC Optimized Settings found it “doesn’t much impact the performance in most scenarios.” I’d still call that a single tester’s result rather than settled fact, but nothing in the other sources contradicts it.
The Few Places the Processor Does Matter
The processor matters at the bottom edge, not in the middle. DSOGaming simulated a dual-core CPU without SMT and the game would not run properly, with stutters that lasted up to 13 seconds; with SMT enabled it held more than 60 FPS at all times. That is a simulated configuration, not a retail CPU, but it fits the official floor: Steam lists a Ryzen 5 1600 or Core i5-8400 minimum for 30 FPS at 720p.
If you do have a real CPU-bound game, it looks different. Our Manor Lords settings guide covers one: its frame rate falls as the village population grows, and no graphics slider fixes it. Final Fantasy 16 has no equivalent behaviour in any source I checked.
Why Lowering the Preset Barely Helps
Dropping the whole preset from Ultra to Low gained 27–39% in Notebookcheck’s laptop tests and only 14–19% in DSOGaming’s native 1080p run. The Notebookcheck rows, all at 1920×1080 with upscaling off, show the spread.
| GPU (1080p, FPS) | Low | Medium | High | Ultra | Low vs Ultra |
|---|---|---|---|---|---|
| RTX 4050 Laptop | 54.4 | 49.0 | 42.3 | 39.0 | 1.39× |
| RTX 4060 Laptop | 64.7 | 61.7 | 53.8 | 48.8 | 1.33× |
| RTX 4070 Laptop | 73.8 | 70.8 | 61.8 | 55.4 | 1.33× |
| RTX 4080 Laptop | 102.4 | 94.6 | 89.5 | 80.3 | 1.27× |
| RTX 4090 Laptop | 129.7 | 123.4 | 109.4 | 98.5 | 1.32× |
The ratio column is my arithmetic on Notebookcheck’s figures. It says a full preset drop is worth roughly a quarter to two-fifths more frames on these cards, and the gap between Low and Medium is small: 4–11% on every row above. Dropping from Ultra to High alone is worth only 8–12% on these rows.
My reading of why, flagged as inference rather than something any source measured: most of the frame cost sits in drawing the image at your output resolution, and presets mostly change effects layered on top of it. That is also why the one lever that changes the pixel count, upscaling, outperforms the whole menu below it.
Setting-by-Setting: What Each One Costs
Only a few settings are worth touching, and Screen Space Reflections and Clutter Density are the two with measurable payoffs. The measured column comes from PC Optimized Settings’ RTX 4090 run unless noted.
| Setting | Measured cost | My pick |
|---|---|---|
| Screen Space Reflections | Lowers average FPS by nearly 10% | Lower it if you need frames |
| Clutter Density | Low is up to 14% faster than High in forested areas | Low if you need frames |
| Graphical Fidelity | “Quite low” tax (pre-1.02); patch 1.02 tuned it for performance | Medium |
| Shadow Quality | “Modest” (pre-1.02); patch 1.02 added clouds to Low and Medium | Medium |
| Terrain Quality | Minimal, near zero in many starting areas | High |
| Water Quality | Nominal | High |
| Ambient Occlusion | About 1 FPS | Leave on |
| Bloom | About 1 FPS | Leave on |
| NPC Density | Little impact in most scenarios | Leave as is |
| Variable Rate Shading | “Indetectable” impact on quality or performance | Off |
| Textures | Not measured for FPS; mainly a VRAM question (see below) | High on 8 GB+ cards |
One caution on the first four rows. Shadow Quality and Graphical Fidelity were measured before patch 1.02, which Square Enix says should “improve performance when low or medium is selected.” Those two may be cheaper to lower now than the table suggests, but I found no post-patch per-setting test to confirm it.
Textures and the 8 GB VRAM Question
Steam’s own requirements list 8 GB of VRAM or above for both minimum and recommended specs, so 8 GB cards sit exactly on the line. The two sources that measured usage disagree. GameGPU recorded 7–8 GB at 1080p and 8–10 GB at 4K, while PC Optimized Settings reported over 13 GB at 4K and 11 GB at 1440p and 1080p. I couldn’t tell which tool or metric each used. If you have 8 GB, start with Textures on High and drop one step only if you see texture pop-in or hitching.
Upscaling and Frame Generation: Where the Frames Are
Quality-mode upscaling gained 45% at 4K on an RTX 4090 in PC Optimized Settings’ test, and Performance mode gained 85%. The full set of upscaling and frame-generation results I found:
| Test | Rig | Result |
|---|---|---|
| PC Optimized Settings, 4K, DLSS | RTX 4090 | 55 FPS native; +45% Quality, +68% Balanced, +85% Performance (up to about 100 FPS) |
| PC Optimized Settings, 4K, frame generation only | RTX 4090 | +64% average FPS |
| Notebookcheck, 1440p Ultra, DLSS Quality | RTX 4090 Laptop | 69.6 FPS native to 97.1 FPS (+40%) |
| GGRecon, 3440×1440, hub area | RTX 3080, i9-10900K | 40 FPS native High; 65 FPS DLSS Balanced; 111 FPS FSR 3 with Frame Generation |
| DSOGaming, native 4K Max | RTX 4090 | Dips into the 50s native; over 80 FPS at all times with DLSS Frame Generation and DLAA |
Even the weakest result in the table, +40%, beats the Ultra-to-Low preset drop on the same kind of card. The comparison is rough because the rigs and resolutions differ, but the direction is consistent in every source.
What You Give Up
An upscaler renders fewer pixels and reconstructs the rest, so it trades some sharpness for frames. Real upscalers use motion data and look much better than a plain resize, but the trade is the same. The softness shows most on fine detail such as foliage and hair, and more in Performance mode than in Quality, so as a general guideline step down one mode at a time instead of jumping to Performance.
Frame generation is a different tool. It inserts extra frames rather than rendering more real ones, so a rule of thumb many testers use is to switch it on only when you already have a playable base frame rate. DSOGaming noticed no input latency or artifacts in its run, but that is one tester on an RTX 4090, and a slower card starting from a lower base may feel different. On a rig that doesn’t reach 60 FPS without it, fix the base with upscaling first.
Which Settings Fit Your Card
The right target depends on your resolution and card, not on a preset name. Use the row closest to your hardware.
| If you have… | Do this | Evidence |
|---|---|---|
| RTX 3070 or RX 6700 XT class, 1080p | Medium fidelity and shadows; add upscaler Quality only if 60 FPS doesn’t hold | GameGPU: 60+ FPS at 1080p Max needs about this tier |
| RTX 4070 Ti or RX 6800 XT class, 1440p | Native is viable at 60; use upscaler Quality for high-refresh monitors | GameGPU: 60+ FPS at 1440p Max needs about this tier |
| RTX 4090, 4K | Upscaler Quality plus frame generation; native Max can’t hold 60 | DSOGaming: 50s–70s native; PC Optimized: 55 FPS native |
| RTX 4060 or 4070 Laptop, 1080p | Upscaler Quality and Medium settings; native Ultra sits under 60 | Notebookcheck: Ultra 48.8 and 55.4 FPS native |
| Ryzen 5 1600 or similar older CPU | Keep SMT on; expect the CPU to matter more here | DSOGaming SMT test; Steam minimum spec |
If you only remember one row, remember the laptop one. An RTX 4060 Laptop runs Ultra at 48.8 FPS and Low at 64.7 in Notebookcheck’s test, so the preset alone gets you over 60 only by dropping everything. Upscaling is the cheaper route to the same number.
How to Check Whether Your Own PC Is GPU-Bound
You can test this in about 30 seconds. Stand somewhere busy, note your frame rate and your GPU utilization, then step the upscaler down one level (or the output resolution down one notch) and look again. If the frame rate rises, you were GPU-bound, which is the expected result for this game. If it barely moves and GPU utilization sits well below its usual ceiling, something else is the limit. As a rough guideline, most people treat sustained utilization in the high 90s as a sign the GPU is the bottleneck.
Smooth average FPS with ugly hitching is a separate problem, and not a CPU or GPU verdict. Reports in the sources match that: GGRecon noted brief stutter on alpha effects such as elemental sparks, and Notebookcheck noted stutter during video sequences regardless of hardware. If your counter looks fine and the picture still catches, our guide to diagnosing frame pacing versus raw FPS explains how to tell frame-time spikes from low frame rate.
Where the Sources Disagree
- How much the preset is worth. DSOGaming measured 14–19% from Max to Low at native 1080p; Notebookcheck’s laptop rows imply 27–39% from Ultra to Low. I trust the direction (small compared with upscaling) more than either figure, since the GPUs and presets differ.
- VRAM use. 7–10 GB (GameGPU) versus 11–13 GB (PC Optimized Settings). Unresolved above.
- What limited Digital Foundry’s mainstream rig. The Ryzen 3600 and RTX 2070 Super at 1440p dropped frames in demanding scenes, and the summary I could read doesn’t say whether the CPU or GPU was the limit. I treat it as GPU-side, consistent with Digital Foundry’s own “GPU-side” description, but that is my reading.
FAQ
Should I lower graphics settings to fix low FPS in Final Fantasy 16?
Not first. Lowering the whole preset gained 14–19% in one test and 27–39% in another, while Quality-mode upscaling gained 40–45% in the 4K and 1440p tests on an RTX 4090 and an RTX 4090 Laptop. Start with upscaling, then lower only Screen Space Reflections and Clutter Density.
Is Final Fantasy 16 CPU intensive on PC?
Not on a reasonable CPU. It scales across many threads, but every test that measured it found the graphics card was the limit, including a 1080p run on a Ryzen 9 7950X. The exception is very low thread counts, where DSOGaming’s simulated dual-core without SMT stuttered badly.
Does NPC Density affect FPS in Final Fantasy 16?
Barely, according to the one test that isolated it. PC Optimized Settings found it did not much affect performance in most scenarios, so lowering it risks emptier-looking towns for little gain.
Is Frame Generation worth using in Final Fantasy 16?
Yes, once your base frame rate is already reasonable. It added 64% at 4K in one test, and an FSR 3 plus Frame Generation run on an RTX 3080 reached 111 FPS against 65 for DLSS Balanced in another. It doesn’t fix a low base frame rate, and slower cards may feel more latency than the RTX 4090 testers reported.
Can an 8 GB graphics card run Final Fantasy 16?
Yes, at its stated floor. Steam lists 8 GB of VRAM or above for both minimum and recommended specs. Keep Textures on High, use an upscaler, and watch for pop-in, since the sources disagree on how much VRAM the game really uses.
Sources
- PC Optimized Settings. Final Fantasy 16 PC Optimization: Best Graphics Settings with Benchmarks. PC Optimized Settings, 18 September 2024
- DSOGaming. Final Fantasy XVI PC Performance Analysis. DSOGaming, 23 September 2024
- DSOGaming. Final Fantasy 16 cannot run with 60fps at native 4K/Max on the NVIDIA RTX 4090. DSOGaming, 16 September 2024
- GGRecon. Final Fantasy XVI PC review. GGRecon
- GameGPU. Final Fantasy XVI PC performance: graphics card and processor benchmarks. GameGPU, 17 September 2024
- Notebookcheck. Final Fantasy XVI im Test: Laptop- und Desktop-Benchmarks. Notebookcheck, 5 October 2024 (updated 18 February 2026)
- Multiplayer.it. Final Fantasy 16 PC: bello da vedere ma pesante, Digital Foundry (settaggi ottimizzati). Multiplayer.it
- Square Enix. FINAL FANTASY XVI Version Update 1.02. Steam News, 17 October 2024
- Square Enix. FINAL FANTASY XVI Version Update 1.03. Steam News, 4 March 2025
- Square Enix. FINAL FANTASY XVI system requirements. Steam
Images in this article are AI illustrations or composites, not game screenshots. Final Fantasy is a trademark of Square Enix; all product names are used for identification only.
I've been playing video games for over 20 years, spanning everything from early PC titles to modern open-world games. I started Switchblade Gaming to publish the kind of accurate, well-researched guides I always wanted to find — built on primary sources, tested in-game, and kept up to date after patches. I currently focus on Minecraft and Pokémon GO.