AMD Radeon RX 9070 XT vs NVIDIA GeForce GTX 1070 Ti Comparison

NVIDIA’s GeForce GTX 1070 Ti and AMD’s Radeon RX 9070 XT are separated by nearly a decade of GPU architecture. The benchmark data shows a generational chasm, with the RX 9070 XT claiming 9 of 10 head-to-head tests. However, one surprising result flips the narrative, making this comparison less one-sided than raw averages suggest.

Head-to-Head Benchmarks

The most decisive victory for the AMD Radeon RX 9070 XT comes in the 3DMark Steel Nomad DX12 test, where it scores 7,260 versus the GTX 1070 Ti’s 1,471. That is a delta of -79.7%, meaning the RX 9070 XT delivers nearly five times the performance in this modern API workload. This is the single largest performance gap in the entire dataset, underscoring how far GPU compute has advanced.

The RX 9070 XT also dominates legacy DirectX tests. In PassMark DirectX 11, it scores 292 against 107, a 63.4% advantage. DirectX 10 shows a 41.6% lead (154 vs. 90), and DirectX 9 shows a 41.4% lead (353 vs. 207). These are not marginal improvements; they represent a consistent doubling or near-doubling of frame throughput in older APIs, which suggests the newer architecture handles draw calls and state changes far more efficiently.

Compute-heavy workloads paint a similar picture. The PassMark GPU Compute test awards the RX 9070 XT 15,864 points versus 7,165 for the GTX 1070 Ti, a 54.8% win. The overall PassMark G3D score also favors AMD heavily: 26,795 vs. 14,674, a 45.2% gap. In DirectX 12, the margin narrows to 34.6% (78 vs. 51), but the RX 9070 XT still wins comfortably.

The AMD card also takes the 2D graphics test, scoring 1,328 versus 876, a 34% lead. Even in Vulkan, where the GTX 1070 Ti is relatively strong, the RX 9070 XT wins 65,879 to 61,006, though the delta is just 7.4%. This is the closest contest between the two cards, indicating that Pascal’s Vulkan driver implementation remains competitive.

However, there is one glaring anomaly. In Geekbench OpenCL, the NVIDIA GeForce GTX 1070 Ti scores 49,219, while the AMD Radeon RX 9070 XT scores only 17,428. That is a 182.4% win for the older card — a staggering reversal. This result is likely an artifact of workload characteristics or driver optimization, but the data is unambiguous: the GTX 1070 Ti crushes the RX 9070 XT in this specific OpenCL benchmark.

The head-to-head record is 9-1 in favor of the RX 9070 XT. Yet the Geekbench OpenCL result prevents a clean sweep, and it highlights that synthetic benchmarks can produce outlier results that contradict the broader trend.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 1070 Ti has an average benchmark score of 14,277, while the AMD Radeon RX 9070 XT averages 13,543. Despite winning 9 of 10 head-to-head tests, the RX 9070 XT’s average is pulled down by its extremely low Geekbench OpenCL score.

Q: How do the two cards compare in 3DMark Steel Nomad?

A: The AMD Radeon RX 9070 XT scores 7,260, which is 79.7% higher than the GTX 1070 Ti’s 1,471. This is the largest performance delta in the head-to-head data.

Q: What is the percentile ranking for each GPU?

A: The GTX 1070 Ti sits at the 56th percentile of all GPUs, while the RX 9070 XT ranks at the 55th percentile. These rankings are nearly identical, despite the RX 9070 XT’s superior raw performance in most tests.

Q: Which card wins in Geekbench Vulkan?

A: The AMD Radeon RX 9070 XT wins with a score of 65,879 versus 61,006 for the GTX 1070 Ti, a 7.4% advantage. This is the smallest margin of victory for AMD in the head-to-head tests.

Q: Does the GTX 1070 Ti win any benchmark?

A: Yes, the GTX 1070 Ti wins the Geekbench OpenCL test with a score of 49,219, which is 182.4% higher than the RX 9070 XT’s 17,428. This is the only head-to-head win for NVIDIA.

Q: What are the nearest rivals for each card?

A: The GTX 1070 Ti’s closest rival is the Intel Iris Xe MAX Graphics, with an average score of 14,315 and a delta of -0.3%. The RX 9070 XT’s nearest rival is the AMD Radeon HD 7770M, averaging 13,536 with a delta of 0.1%.

Where Each One Wins

The AMD Radeon RX 9070 XT is the clear winner for modern gaming workloads. Its 3DMark Steel Nomad DX12 score of 7,260 indicates strong performance in current-generation APIs, and its PassMark DirectX 12 score (78 vs. 51) reinforces this. For DX11 and DX10 titles, the RX 9070 XT’s 292 and 154 scores respectively suggest it can handle legacy games at much higher settings and frame rates. The GPU compute score of 15,864 also makes it the choice for tasks like rendering, physics simulation, and machine learning inference.

The NVIDIA GeForce GTX 1070 Ti has only one standout use case: OpenCL compute. Its 49,219 Geekbench OpenCL score is more than double that of the RX 9070 XT, meaning applications that rely heavily on OpenCL acceleration — such as certain video encoding tools or scientific software — would run significantly faster on the older card. This is a niche advantage, but for users locked into OpenCL-based workflows, the GTX 1070 Ti remains surprisingly relevant.

In 2D desktop workloads, the RX 9070 XT’s PassMark G2D score of 1,328 versus 876 suggests snappier window management and higher-resolution desktop rendering. For Vulkan-based titles, the RX 9070 XT wins but only by 7.4%, so the GTX 1070 Ti can still hold its own in Vulkan games, particularly at lower resolutions where the bottleneck shifts to the CPU.

Specification Differences

Memory is a major differentiator. The RX 9070 XT has 16 GB of GDDR6 memory with a bandwidth of 644.6 GB/s, while the GTX 1070 Ti has 8 GB of GDDR5 memory at 256.3 GB/s. Both use a 256-bit bus, but the newer memory type and higher effective speed (20.1 Gbps vs. 8 Gbps) give AMD a 2.5x bandwidth advantage.

Compute resources are similarly lopsided. The RX 9070 XT packs 4,096 shading units, 256 TMUs, and 128 ROPs, versus 2,432 shading units, 152 TMUs, and 64 ROPs on the GTX 1070 Ti. The RX 9070 XT also has 64 ray tracing cores, a feature entirely absent from the Pascal-based GTX 1070 Ti. Pixel rate (380.2 vs. 107.7 GPixel/s) and texture rate (760.3 vs. 255.8 GTexel/s) both favor AMD by roughly 3x.

Clock speeds differ substantially. The RX 9070 XT boosts to 2,970 MHz with a game clock of 2,400 MHz, while the GTX 1070 Ti boosts to only 1,683 MHz. The RX 9070 XT also has a higher base clock (1,660 MHz vs. 1,607 MHz). Power requirements scale accordingly: the RX 9070 XT has a 304 W TDP and needs a 700 W PSU with two 8-pin connectors, whereas the GTX 1070 Ti draws 180 W and requires a 450 W PSU with a single 8-pin connector.

Interface and outputs also differ. The RX 9070 XT uses PCIe 5.0 x16 and offers 1x HDMI 2.1b plus 3x DisplayPort 2.1a. The GTX 1070 Ti uses PCIe 3.0 x16 and has 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.4a. The RX 9070 XT supports DirectX 12 Ultimate (12_2), while the GTX 1070 Ti is limited to DirectX 12 (12_1).

Architecture Differences

The RX 9070 XT is built on TSMC’s 4 nm process node with RDNA 4.0 architecture, while the GTX 1070 Ti uses TSMC’s 16 nm node with Pascal architecture. This process shrink allows AMD to pack 53,900 million transistors into a 357 mm² die, yielding a transistor density of 151.0M per mm². The GTX 1070 Ti contains 7,200 million transistors on a 314 mm² die, a density of just 22.9M per mm². That is a 6.6x density advantage for the newer chip.

The RX 9070 XT’s RDNA 4.0 design includes dedicated ray tracing hardware, a capability Pascal lacks entirely. It also features 1:1 FP16 to FP32 throughput (48.66 TFLOPS for both), whereas the GTX 1070 Ti has severely crippled FP16 performance at 127.9 GFLOPS (a 1:64 ratio). This makes the RX 9070 XT far more capable in mixed-precision workloads.

The RX 9070 XT is manufactured as an active product, released on 2025-03-05, succeeding the Navi III generation. The GTX 1070 Ti is end-of-life, released on 2017-11-01, with the GeForce 20 series as its successor. The RX 9070 XT’s launch MSRP is 599 USD, while the GTX 1070 Ti launched at 399 USD. The RX 9070 XT also supports DirectX 12 Ultimate and Vulkan 1.4, matching the GTX 1070 Ti’s Vulkan 1.4 support, but adds newer display outputs and a PCIe 5.0 interface.