AMD Radeon RX 6700S vs NVIDIA GeForce RTX 3070 Mobile Comparison

The AMD Radeon RX 6700S and NVIDIA GeForce RTX 3070 Mobile represent two distinct approaches to high-performance mobile graphics, with the data showing a decisive victory for the NVIDIA part in the majority of modern workloads. Across the ten head-to-head benchmark comparisons, the RTX 3070 Mobile secures eight wins, while the RX 6700S manages two. The NVIDIA GPU’s average benchmark score of 20,534 places it at the 65th percentile among all GPUs, while the AMD part’s average of 22,154 sits at the 67th percentile, a slight inversion that underscores how legacy and compute tests can skew aggregate figures. The RTX 3070 Mobile’s dominance is most pronounced in newer DirectX 12 workloads and compute tasks, whereas the RX 6700S shows its strengths in specific legacy DirectX 9 tests and 2D operations. This analysis will break down the architectural underpinnings, benchmark deltas, and use-case implications of these two end-of-life mobile GPUs.

The Verdict

The benchmark data paints a clear picture: the NVIDIA GeForce RTX 3070 Mobile is the superior choice for users prioritizing modern gaming performance and GPU compute throughput. Its lead in the 3DMark Steel Nomad DX12 test is substantial, scoring 2,380 against the RX 6700S’s 1,451, a 39% advantage that indicates significantly stronger DirectX 12 gaming capability. The RTX 3070 Mobile also delivers a 23.6% lead in Geekbench OpenCL (92,939 vs. 70,965) and a 12.3% advantage in Passmark GPU compute (6,827 vs. 5,990), making it the better option for content creation, rendering, and any workflow leveraging general-purpose GPU compute. For those who primarily play modern titles or use GPU-accelerated applications, the RTX 3070 Mobile is the data-backed pick.

The AMD Radeon RX 6700S, however, is not without merit. It wins the Passmark DirectX 9 test decisively, scoring 195 versus 160, a 21.9% margin, and the Passmark G2D test with 739 versus 641, a 15.3% lead. This suggests that for users with legacy software libraries, older game titles running on DirectX 9, or workloads heavily dependent on 2D graphics operations, the RX 6700S offers a tangible performance advantage. Its aggregate percentile ranking, at 67 vs. 65, is also slightly higher, driven by its strong showing in Vulkan (76,150 vs. 86,768, a 12.2% loss) and its closer margin in Passmark G3D (15,066 vs. 15,309, a 1.6% loss). If the workload is predominantly 2D or relies on DirectX 9, the RX 6700S is the rational choice, but for almost everything else, the RTX 3070 Mobile leads.

Architecture Differences

The two GPUs are built on fundamentally different architectures and process nodes. The AMD Radeon RX 6700S utilizes the RDNA 2.0 architecture on a 7 nm TSMC process, featuring the Navi 23 chip with 11,060 million transistors on a 237 mm² die, resulting in a transistor density of 46.7M per mm². In contrast, the NVIDIA GeForce RTX 3070 Mobile uses the Ampere architecture on an 8 nm Samsung process, with the GA104 chip housing 17,400 million transistors on a 392 mm² die, yielding a lower density of 44.4M per mm². This means the NVIDIA chip is physically larger and denser in absolute transistor count, while the AMD chip benefits from a more advanced process node.

The compute configurations diverge significantly. The RX 6700S has 1,792 shading units, 112 texture mapping units (TMUs), and 64 raster operation units (ROPs), complemented by 28 ray tracing cores. The RTX 3070 Mobile features a much higher count of 5,120 shading units, 160 TMUs, and 80 ROPs, plus 40 RT cores and 160 tensor cores, which are absent from the AMD part. This massive shader count disparity explains the NVIDIA GPU’s FP32 throughput of 15.97 TFLOPS versus the AMD part’s 7.168 TFLOPS, a 2.2x advantage. The RTX 3070 Mobile’s FP16 performance is identical at 15.97 TFLOPS (1:1), while the RX 6700S achieves 14.34 TFLOPS via a 2:1 ratio, indicating a different approach to half-precision compute.

Memory subsystems also differ. Both GPUs have 8 GB of GDDR6 memory, but the RTX 3070 Mobile uses a 256-bit bus to deliver 448.0 GB/s of bandwidth, while the RX 6700S is limited to a 128-bit bus and 224.0 GB/s. The clock speeds reflect their power envelopes: the RX 6700S has a base clock of 1700 MHz and boost of 2000 MHz, while the RTX 3070 Mobile runs lower at 1110 MHz base and 1560 MHz boost, though its higher core count compensates. The RX 6700S is rated at a TDP of 80 W and is an IGP (integrated) form factor with no power connectors, whereas the RTX 3070 Mobile has a 115 W TDP, also with no external power connectors. The bus interface also differs, with the AMD part using PCIe 4.0 x8 and the NVIDIA part using PCIe 4.0 x16. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, but the RTX 3070 Mobile’s tensor cores give it a distinct advantage in AI-accelerated workloads.

Head-to-Head Benchmarks

The largest single benchmark delta in the head-to-head comparison is in the 3DMark Steel Nomad DX12 test. The RTX 3070 Mobile scores 2,380, while the RX 6700S scores 1,451, resulting in a 39% lead for NVIDIA. This is a massive gap that speaks to the NVIDIA part’s superior raw rasterization and DX12 feature utilization. This is the most modern gaming workload in the list, and the RTX 3070 Mobile’s 5120 shading units versus 1792 for the AMD part are the primary driver of this result.

In compute-centric benchmarks, the RTX 3070 Mobile again leads. Geekbench OpenCL shows a score of 92,939 vs. 70,965, a 23.6% margin. Passmark GPU compute shows 6,827 vs. 5,990, a 12.3% lead. These results are consistent with the FP32 throughput difference, where the NVIDIA part’s 15.97 TFLOPS exceeds the AMD part’s 7.168 TFLOPS. The Vulkan test is closer but still favors NVIDIA, with 86,768 vs. 76,150, a 12.2% delta, suggesting that the RX 6700S’s RDNA 2.0 architecture handles Vulkan relatively well but cannot overcome the core count deficit.

The AMD Radeon RX 6700S’s wins are concentrated in older or less GPU-intensive tests. In Passmark DirectX 9, it scores 195 against 160, a 21.9% advantage. This is a notable outlier, indicating that the RDNA 2.0 architecture retains strong performance in legacy DirectX 9 workloads, possibly due to driver optimizations or architectural efficiency in that API. The other win is in Passmark G2D, where it scores 739 vs. 641, a 15.3% lead. This 2D graphics test measures basic desktop and image manipulation performance, where the AMD part’s higher base clock of 1700 MHz likely contributes to its advantage.

The remaining benchmarks are closer but still favor NVIDIA. Passmark DirectX 11 shows 138 vs. 134, a narrow 2.9% margin, while Passmark DirectX 12 shows 64 vs. 58, a 9.4% delta. Passmark G3D, a general 3D performance metric, shows 15,309 vs. 15,066, only a 1.6% difference. These results indicate that in synthetic 3D workloads without heavy DX12 or compute requirements, the two GPUs are much closer, with the RX 6700S often within striking distance. The overall win count of 8-2 in favor of NVIDIA, however, reflects the RTX 3070 Mobile’s broader dominance.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon RX 6700S has an average benchmark score of 22,154, which is higher than the NVIDIA GeForce RTX 3070 Mobile’s average of 20,534. This places the AMD part at the 67th percentile versus the NVIDIA part’s 65th percentile.

Q: What is the biggest performance gap in the head-to-head tests?

A: The largest delta is in the 3DMark Steel Nomad DX12 test, where the NVIDIA GeForce RTX 3070 Mobile scores 2,380 and the AMD Radeon RX 6700S scores 1,451, giving NVIDIA a 39% lead.

Q: In which benchmark does the AMD Radeon RX 6700S have its largest win?

A: The AMD Radeon RX 6700S’s largest win is in the Passmark DirectX 9 test, where it scores 195 versus the RTX 3070 Mobile’s 160, a 21.9% advantage.

Q: How do the two GPUs compare in memory bandwidth?

A: The NVIDIA GeForce RTX 3070 Mobile has a 256-bit memory bus providing 448.0 GB/s of bandwidth, while the AMD Radeon RX 6700S has a 128-bit bus providing 224.0 GB/s. Both have 8 GB of GDDR6 memory.

Q: What is the difference in FP32 compute performance?

A: The NVIDIA GeForce RTX 3070 Mobile delivers 15.97 TFLOPS of FP32 performance, while the AMD Radeon RX 6700S delivers 7.168 TFLOPS, making the NVIDIA part more than twice as fast in this metric.

Q: Which GPU has more shading units?

A: The NVIDIA GeForce RTX 3070 Mobile has 5,120 shading units, while the AMD Radeon RX 6700S has 1,792 shading units. The NVIDIA part also has 160 tensor cores, which the AMD part lacks entirely.

Where Each One Wins

The NVIDIA GeForce RTX 3070 Mobile is the winner in all modern gaming and compute scenarios. Its 39% lead in 3DMark Steel Nomad DX12 makes it the clear choice for DirectX 12 titles, which are the standard for current AAA games. The 23.6% lead in Geekbench OpenCL and 12.3% lead in Passmark GPU compute establish it as the superior option for rendering, video encoding, machine learning inference (thanks to its 160 tensor cores), and any workload that can leverage its 15.97 TFLOPS of FP32 throughput. The RTX 3070 Mobile’s 448.0 GB/s of memory bandwidth, double that of the RX 6700S, also benefits high-resolution textures and data-intensive tasks. For users playing newer games, doing 3D modeling, or using GPU-accelerated applications, the RTX 3070 Mobile is the definitive pick, despite its 115 W TDP being higher than the RX 6700S’s 80 W.

The AMD Radeon RX 6700S wins in specific legacy and 2D scenarios. Its 21.9% advantage in Passmark DirectX 9 makes it the better choice for older game libraries that rely on that API, such as many titles from the early 2000s or indie games with DX9 renderers. The 15.3% win in Passmark G2D indicates superior 2D graphics performance, which could benefit users who work with large image files, do desktop applications with high-resolution displays, or run software that is not GPU-accelerated in the traditional 3D sense. The RX 6700S also has a higher base clock of 1700 MHz versus 1110 MHz, which contributes to its responsiveness in these lighter workloads. For users with a specific need for DX9 compatibility or 2D performance, the RX 6700S’s wins make it a viable, lower-power alternative, but its utility is otherwise limited to a narrow set of use cases.