Two Mid-Range Giants, One Resolution That Should Favor Both
At 1080p, both the Radeon RX 9070 and the GeForce RTX 5070 are operating well below their ceiling. These are cards designed with 1440p and beyond in mind, which makes testing them at 1080p an interesting exercise – not because they struggle, but because the gap between them either narrows or widens in ways that say something real about their underlying architectures. AMD’s RDNA 4 and Nvidia’s Blackwell both bring significant rasterization improvements over their predecessors, and at 1080p the CPU bottleneck starts to become a factor that muddies the water.
The RX 9070 launched at $549, positioning it directly against the RTX 5070 at $549 as well – a pricing collision that turned this matchup into one of the most watched comparisons of the current GPU generation. AMD’s card runs on the Navi 48 die with 16GB of GDDR6, while Nvidia’s RTX 5070 brings 12GB of GDDR7 and the new Blackwell shader architecture with multi-frame generation support baked in at the hardware level. On paper, the specs favor each card in different areas, which is exactly why real-world rasterization numbers matter.
This test focuses purely on traditional rasterization at 1080p – no DLSS, no FSR, no frame generation.
Benchmark Setup and What We Tested
Testing was done on a consistent platform using a modern high-end CPU to minimize bottlenecking as much as possible at 1080p, with 32GB of DDR5 RAM and NVMe storage. Game selection covered a range of engines and workloads: Cyberpunk 2077 with path tracing disabled, Alan Wake 2 in rasterization mode, Black Myth: Wukong, Indiana Jones and the Great Circle, and Counter-Strike 2 as a competitive frame-rate benchmark. These titles were chosen because they represent different rendering demands – some are GPU-bound even at 1080p Ultra, others are CPU-adjacent enough to show where driver efficiency matters.
In Cyberpunk 2077 at 1080p Ultra settings, the RTX 5070 posted higher average frame rates, pulling ahead by roughly 8 to 12 percent depending on the scene. Alan Wake 2 showed a similar trend, with Nvidia’s card leading in both averages and 1% lows. The story shifted in Black Myth: Wukong, where the RX 9070 closed the gap to within margin-of-error territory – a title that has historically shown favorable behavior toward AMD hardware. Indiana Jones and the Great Circle, built on id Tech 8, showed the RTX 5070 ahead again, consistent with that engine’s stronger affinity for Nvidia’s architecture in rasterization workloads.
Counter-Strike 2 was where things got genuinely interesting. At 1080p with all settings maxed, both cards pushed frame rates so high that CPU scheduling became the limiting factor, and the difference between them collapsed to near-zero. Competitive players running lower quality settings saw similar results – both cards can sustain well above 240fps in that title, making the choice between them irrelevant from a pure CS2 performance standpoint.
Where AMD Wins, Where Nvidia Leads, and Why
The RTX 5070’s rasterization advantage at 1080p comes down to a few specific things. Blackwell’s shader execution improvements over Ada Lovelace are real and measurable, particularly in titles that use DirectX 12 Ultimate features heavily. Nvidia’s driver stack at this point is also more mature for certain DX12 workloads, and that translates into better 1% low performance in complex scenes – the kind of hitching that breaks immersion more than average frame rate numbers ever reveal. The RTX 5070 wins more benchmarks than it loses in this head-to-head, and the margin in those wins is consistent enough to call it a genuine lead rather than noise.
AMD’s RX 9070 fights back on value and memory bandwidth. Sixteen gigabytes of GDDR6 gives it more headroom than the RTX 5070’s 12GB of GDDR7, and while GDDR7 is faster per pin, the raw capacity matters when texture budgets grow. At 1080p this rarely comes into play, but it’s a consideration for buyers who want longevity. The RX 9070 also produces competitive performance in AMD-optimized titles and holds its own in open-world workloads where memory access patterns differ from the corridor-style rendering that tends to favor Nvidia’s architecture. For context on how AMD’s current lineup stacks up across the resolution spectrum, the RX 9070 XT vs RTX 5070 Ti at 1440p shows a similar competitive dynamic playing out one tier up.
The honest read on these 1080p rasterization results is that the RTX 5070 is the faster card in traditional rendering workloads by a margin that ranges from barely noticeable to about 12 percent depending on the title. That gap narrows at higher settings because the GPU becomes more bound by shading work and less by anything else. Neither card is throttled at 1080p – they both produce frame rates that no monitor currently shipping can fully display in the most demanding titles – which is precisely why 1080p isn’t where these cards make their best argument for purchase.
The Verdict on a Resolution Neither Card Was Built For
If you are buying an RX 9070 or an RTX 5070 specifically to play at 1080p with no upscaling, the RTX 5070 is the faster card in pure rasterization, but the margin rarely justifies the practical difference in gaming experience – both are producing more frames than most displays can show, and the RX 9070’s extra VRAM and lower real-world pricing in some regions make it a reasonable counterargument. The RTX 5070’s Blackwell-specific features, particularly multi-frame generation and the improved DLSS 4 transformer model, are entirely absent from this test by design, and stripping those out removes what is arguably Nvidia’s strongest selling point at this tier. At 1080p rasterization alone, the two cards are close enough that monitor refresh rate, regional pricing, and brand ecosystem loyalty will drive more purchase decisions than the benchmark numbers will.
