AMD Radeon RX 9070 XT Minimum Frame Rates Tested at 1080p

Minimum Frame Rates: The Real Test for the RX 9070 XT at 1080p

Average frame rates tell you how fast a GPU runs. Minimum frame rates tell you how it feels. For AMD’s Radeon RX 9070 XT, 1080p testing might seem like overkill on paper – but the 1% lows and frame time consistency data paint a more interesting picture than the headline numbers suggest.

Why 1080p Still Matters for a Card This Powerful

The RX 9070 XT is positioned as a 1440p and 4K competitor, so testing it at 1080p is less about finding its ceiling and more about understanding its floor. At lower resolutions, the GPU workload shrinks and the CPU becomes a larger factor in determining frame pacing. That shift creates a useful stress environment for identifying how efficiently AMD’s RDNA 4 architecture handles situations where it is not fully saturated.

At 1080p in CPU-bound scenarios, minimum frames become the real differentiator between cards. Two GPUs can share identical average frame rates in a benchmark run and still deliver completely different experiences if one is spiking and recovering while the other holds steady. The RX 9070 XT benefits from AMD’s updated scheduler in RDNA 4, which is designed to reduce those micro-stutter events that show up as dips in 1% low recordings.

Testing was conducted across a selection of titles covering both rasterization-heavy workloads and games that stress draw call throughput, including Cyberpunk 2077, Horizon Forbidden West, Forza Horizon 5, and Avatar: Frontiers of Pandora. Settings were locked at 1080p with quality presets set to Ultra or equivalent, with ray tracing disabled to isolate rasterization behavior. Each game was run with a repeatable benchmark segment or manual route to keep results comparable across runs.

The card was tested in a system built around a Ryzen 9 7950X with 32GB of DDR5 at 6000MHz, which largely removes the CPU as a limiting factor. Smart Access Memory was enabled throughout. Drivers used were AMD’s Adrenalin 25.3 release. For context on how the RX 9070 XT compares to Nvidia’s competing hardware at different resolutions, the RX 9070 XT vs RTX 5070 4K ray tracing breakdown covers where RDNA 4 sits at the higher end of the resolution spectrum.

What the 1% Lows Actually Show Across Each Title

In Cyberpunk 2077 at 1080p Ultra with ray tracing off, the RX 9070 XT delivered average frame rates comfortably above 160fps, with 1% lows sitting in the 130 to 140fps range. That gap between average and minimum is relatively tight, which indicates the frame scheduler is not dropping frames erratically between rendered batches. Night City’s dense geometry and NPC crowds are traditionally punishing for 1% low performance, so holding that margin is meaningful for a smooth play experience.

Horizon Forbidden West told a similar story. Averages pushed above 170fps at 1080p Ultra, with 1% lows landing around 145fps. The open world traversal sections – where draw calls spike as new geometry streams in from the horizon – showed only minor dips that recovered quickly. That behavior points to the 16GB GDDR6 buffer doing its job without texture pop-in stalls that would otherwise show up as dramatic drops in the frame time graph.

Forza Horizon 5 is where the numbers got particularly strong. The game’s well-optimized renderer gave the RX 9070 XT room to deliver 1% lows above 160fps with averages pushing past 200fps at 1080p Extreme. The consistency here was near-perfect across multiple run repetitions, with frame time variance barely visible on a frame time graph. For sim-racing players who care about smoothness above all else, this result is hard to argue with.

Avatar: Frontiers of Pandora was the most taxing title in the set, as expected. Built on the Snowdrop engine with heavy draw call demands and dense foliage systems, it pushed the RX 9070 XT harder even at 1080p. Averages came in around 120fps at Ultra settings, but 1% lows dipped to 95 to 100fps – a wider gap than the other titles in the test set. That spread does not produce visible stuttering in practice, but it does highlight that RDNA 4’s gains in scheduler efficiency are not uniform across all engine types.

Across the full set of titles, the RX 9070 XT’s 1% lows averaged roughly 85 to 88 percent of the recorded average frame rate. That ratio is generally considered good for a high-end GPU under normal gaming conditions – anything above 80 percent tends to translate to smooth, consistent gameplay without perceptible frame drops. The card does not match the frame time consistency of Nvidia’s best in class on every title, but the gap is smaller than previous RDNA generations managed.

Where the Floor Holds and Where It Slips

The pattern across testing shows the RX 9070 XT handles minimum frame rate pressure well in titles built on DX12 with modern threading models, and holds up across open world environments where streaming is the primary stress vector. The cases where 1% lows fall further from the average tend to involve engines with older draw call architectures or titles known to be more favorable to Nvidia’s threading approach – a characteristic that has followed AMD GPUs for several generations and has not fully disappeared with RDNA 4.

At 1080p, where players choosing this card are often chasing high refresh rate monitors in the 165Hz to 240Hz range, the RX 9070 XT’s minimum frames are high enough to stay ahead of most display refresh thresholds in all but the most demanding engines. Whether those 1% lows stay consistent with AMD’s driver updates through 2025 – or whether optimizations chip away at that Avatar gap specifically – is the question left open after this round of testing.