AMD Radeon RX 9060 vs NVIDIA GeForce RTX 2070 Comparison

The NVIDIA GeForce RTX 2070 and AMD Radeon RX 9060 represent two very different eras of GPU design, and the benchmark data reflects that divide clearly. The RTX 2070, a Turing-era part, leans on its mature driver stack and specific architectural strengths, while the RX 9060, built on RDNA 4.0, shows a clear preference for modern, compute-heavy workloads. The recorded data shows a split: the RX 9060 wins seven of the ten head-to-head comparisons, but the RTX 2070 secures decisive victories in three specific tests, including a massive one in Vulkan performance. This is not a simple generational sweep; it is a nuanced matchup where the API and workload type dictate the winner.

Where Each One Wins

The AMD Radeon RX 9060 establishes its dominance in modern DirectX 12 and compute-oriented tasks. Its most striking victory comes in the 3DMark Steel Nomad DX12 test, where it scores 3322 against the RTX 2070’s 1569, a delta of 52.8% in AMD’s favor. This is a workload that stresses raw throughput and modern geometry processing, areas where the RX 9060’s newer architecture excels. The RX 9060 also wins in PassMark’s DirectX 11 test (182 vs. 129, a 29.1% lead) and DirectX 9 (280 vs. 211, a 24.6% lead), indicating better performance in legacy DX11 titles and older DX9 games that still have a large installed base.

The RX 9060 further extends its lead in general compute and rasterization. In PassMark GPU Compute, it scores 9919 versus 6411, a 35.4% advantage, showing its strength in non-gaming parallel workloads. Its PassMark G3D score of 17631 is 8.7% higher than the RTX 2070’s 16094, and in the G2D test it leads by 18.3% (1002 vs. 819). Even in Geekbench OpenCL, the RX 9060 wins by 9.3% (88183 vs. 79966), though this is a closer contest than the other compute tests.

The NVIDIA GeForce RTX 2070, however, claims a resounding victory in Geekbench Vulkan, scoring 82521 against the RX 9060’s 39476, a 109% advantage. This is the single largest delta in the entire comparison. The RTX 2070 also wins the PassMark DirectX 12 test (60 vs. 44, a 36.4% lead) and narrowly takes the DirectX 10 test (111 vs. 104, a 6.7% lead). These wins suggest that in certain specific API paths, particularly Vulkan and older DX12 implementations, the RTX 2070’s Turing architecture still holds a significant advantage over the newer RDNA 4.0 part.

The Verdict

The data points to a clear split based on application. For users running modern DirectX 12 titles or compute-heavy applications like rendering and physics simulations, the AMD Radeon RX 9060 is the stronger choice. Its 52.8% lead in 3DMark Steel Nomad and 35.4% lead in GPU compute are not marginal; they represent a substantial performance class difference in these workloads. The RX 9060 also holds a 29.1% edge in DirectX 11, which covers a vast library of current and recent games. Its higher PassMark G3D score (17631 vs. 16094) suggests better overall rasterization performance in typical gaming scenarios.

The NVIDIA GeForce RTX 2070 is the pick for Vulkan-based applications and specific legacy API paths. The 109% advantage in Geekbench Vulkan is extraordinary, and the 36.4% win in PassMark DirectX 12 shows that not all DX12 implementations favor the newer card. Users with a library of Vulkan-native games or applications that rely on that API should consider the RTX 2070. The RTX 2070 also has the distinction of a higher average benchmark score overall (18789 vs. 16014), which is influenced by its stellar Vulkan result and places it in the 63rd percentile of all GPUs, compared to the RX 9060’s 59th percentile.

However, the RX 9060’s wins are more numerous and spread across more common workloads. The RTX 2070 is an end-of-life product, while the RX 9060 is active, and the RX 9060’s performance in the 3DMark Steel Nomad benchmark, a modern DX12 test, is a strong indicator of future-proofing. The verdict is that the RX 9060 is the better all-rounder for current and upcoming software, whereas the RTX 2070 is a specialist that excels in Vulkan and some DX12 scenarios.

Head-to-Head Benchmarks

The most significant win for the AMD Radeon RX 9060 is in 3DMark Steel Nomad DX12. The score of 3322 for AMD versus 1569 for NVIDIA results in a 52.8% delta, the largest margin in the entire dataset. This test is designed for modern hardware, and the RX 9060’s RDNA 4.0 architecture, with its 21.43 TFLOPS FP32 compute, is clearly better suited to it. The RTX 2070’s 7.465 TFLOPS FP32 is simply not competitive in this scenario.

The second-largest win for the RX 9060 is in PassMark GPU Compute, where it scores 9919 versus 6411, a 35.4% lead. This aligns with the RX 9060’s higher shading unit count efficiency and its 1:1 FP16 ratio, which allows it to process more parallel instructions. The RTX 2070’s FP16 is listed at a 2:1 ratio, meaning it is less efficient in those workloads.

The RX 9060 also dominates in the DirectX 11 test, scoring 182 to the RTX 2070’s 129, a 29.1% delta. This is a critical result for gaming, as DX11 remains a common fallback and compatibility layer. The RX 9060’s lead in DirectX 9 (280 vs. 211, 24.6%) and G2D (1002 vs. 819, 18.3%) further shows its strength in legacy and 2D applications. The G3D score of 17631 versus 16094 is an 8.7% win, and the OpenCL score of 88183 versus 79966 is a 9.3% win.

For the NVIDIA GeForce RTX 2070, the headline result is the Geekbench Vulkan score. At 82521 versus 39476, the RTX 2070 is more than double the RX 9060, a 109% delta. This is a massive outlier in the data and suggests that the Turing architecture’s Vulkan driver implementation is exceptionally well-optimized. The RX 9060’s Vulkan score is its weakest result relative to its other benchmarks, which is unusual.

The RTX 2070 also wins the PassMark DirectX 12 test (60 vs. 44, a 36.4% lead). This is interesting because the RX 9060 wins the more modern 3DMark DX12 test, but the older PassMark DX12 test favors the RTX 2070. This indicates that DX12 performance is highly dependent on the specific rendering paths used. The RTX 2070 narrowly takes the DirectX 10 test (111 vs. 104, a 6.7% lead), which is a minor but notable win.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce RTX 2070 has a higher average benchmark score of 18789, compared to the AMD Radeon RX 9060’s 16014. However, the RX 9060 wins more individual head-to-head tests.

Q: Why is the Geekbench Vulkan score so much higher for the RTX 2070?

A: The RTX 2070 scores 82521 in Geekbench Vulkan, which is 109% higher than the RX 9060’s 39476. This is the largest performance delta in the comparison and indicates a significant architectural or driver-level advantage for NVIDIA’s Turing design in this specific API test.

Q: Is the RX 9060 better for modern DirectX 12 games?

A: Yes, the data shows the RX 9060 is substantially better in the 3DMark Steel Nomad DX12 test, scoring 3322 versus 1569, a 52.8% advantage. This suggests it is better optimized for current DX12 rendering techniques.

Q: How do the two cards compare in overall PassMark G3D performance?

A: The RX 9060 leads in PassMark G3D with a score of 17631, which is 8.7% higher than the RTX 2070’s 16094. This indicates better general 3D rendering performance for the AMD card.

Q: Which card is better for compute workloads like OpenCL?

A: The RX 9060 is better, scoring 88183 in Geekbench OpenCL versus the RTX 2070’s 79966, a 9.3% lead. It also has a much larger 35.4% lead in PassMark GPU Compute (9919 vs. 6411).

Q: What is the difference in their DirectX 11 performance?

A: The RX 9060 wins the PassMark DirectX 11 test by a 29.1% margin, scoring 182 against the RTX 2070’s 129. This makes the AMD card the stronger choice for the large library of DX11 games.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The NVIDIA GeForce RTX 2070 uses the TU106 chip, based on the Turing architecture, and is fabricated on a 12 nm process at TSMC. It contains 10,800 million transistors on a 445 mm² die, resulting in a transistor density of 24.3M per mm². In contrast, the AMD Radeon RX 9060 uses the Navi 44 chip, based on RDNA 4.0, and is built on a 4 nm process, also at TSMC. This chip packs 29,700 million transistors into a much smaller 199 mm² die, achieving a far higher density of 149.2M per mm².

The core configurations differ significantly. The RTX 2070 has 2304 shading units, 144 TMUs, and 64 ROPs, along with 36 RT cores and 288 tensor cores. The RX 9060 has fewer shading units (1792) and TMUs (112), but the same 64 ROPs, and it has 28 RT cores. Despite having fewer shading units, the RX 9060 achieves a higher FP32 throughput of 21.43 TFLOPS versus the RTX 2070’s 7.465 TFLOPS, due to its much higher boost clock of 2990 MHz compared to 1620 MHz. The RX 9060 also handles FP16 at a 1:1 ratio (21.43 TFLOPS), while the RTX 2070 does so at a 2:1 ratio (14.93 TFLOPS).

Memory configurations also diverge. Both cards have 8 GB of GDDR6, but the RTX 2070 uses a 256-bit bus with a bandwidth of 448.0 GB/s, while the RX 9060 uses a 128-bit bus with a bandwidth of 288.0 GB/s. The RX 9060 compensates with a higher memory clock of 2250 MHz (18 Gbps effective) versus the RTX 2070’s 1750 MHz (14 Gbps effective). The power profiles are notably different: the RTX 2070 has a TDP of 175 W and a suggested PSU of 450 W, while the RX 9060 is more efficient, with a TDP of 132 W and a suggested PSU of 300 W. Both are dual-slot cards with a single 8-pin power connector.

The bus interface also differs, with the RTX 2070 using PCIe 3.0 x16 and the RX 9060 using the newer PCIe 5.0 x16. Display outputs vary as well: the RTX 2070 offers 1x DVI, 1x HDMI 2.0, 2x DisplayPort 1.4a, and 1x USB Type-C, while the RX 9060 provides 1x HDMI 2.1b and 2x DisplayPort 2.1a. The RTX 2070 is an end-of-life product released in 2018, while the RX 9060 is active and released in 2025. The RTX 2070 has a launch MSRP of 499 USD.