AMD Radeon RX 6700M vs AMD Radeon RX 6700S Comparison

The Verdict

The benchmark data splits cleanly between two very different mobile GPUs that share an RDNA 2.0 foundation but target distinct priorities. The AMD Radeon RX 6700M is the performance leader in most compute and modern graphics workloads, winning 6 of 11 head-to-head tests, including the most demanding DirectX 12 and API-agnostic benchmarks. The AMD Radeon RX 6700S, despite its lower raw specifications, takes 5 wins, mostly in legacy DirectX tests and 2D/compute scenarios where its higher base clock and different memory configuration matter more.

The RX 6700M posts an average benchmark score of 25,633, placing it in the 71st percentile of all GPUs. Its nearest rivals include the AMD Radeon Pro W5700 (delta -0.4%), NVIDIA GeForce RTX 3080 Ti Mobile (delta -0.4%), AMD FirePro D700 (-0.8%), and AMD FirePro W7100 (-0.9%). This puts the 6700M in the upper-midrange of the database, within a fraction of a percent of several desktop-class workstation cards.

The RX 6700S, with an average score of 22,154, sits in the 67th percentile. Its nearest rivals are the AMD Radeon RX 5700 (delta -0.1%), NVIDIA GeForce RTX 5060 Mobile (-1.3%), NVIDIA GeForce GTX 1060 6 GB (+1.4%), and AMD Radeon RX 7700 XT (-1.7%). The RX 6700S is essentially tied with the RX 5700 in average score, while trailing the RTX 5060 Mobile by 1.3%.

For a user choosing between them, the RX 6700M is the clear pick for modern DirectX 12 gaming, Vulkan titles, and Metal-based workloads. The RX 6700S is better suited for users running older DirectX 9/11 applications or prioritizing 2D interface performance and raw compute throughput per watt, given its much lower 80 W power envelope versus 135 W.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon RX 6700M averages 25,633 across all recorded tests, which is 3,479 points higher than the RX 6700S average of 22,154. The RX 6700M sits in the 71st percentile of all GPUs, while the RX 6700S sits in the 67th percentile.

Q: How much faster is the RX 6700M in DirectX 12?

A: In the 3DMark Steel Nomad DX12 test, the RX 6700M scores 1,845 versus 1,451 for the RX 6700S, a 27.2% advantage. In the Passmark DirectX 12 test, the RX 6700M leads 65 to 58, a 12.1% difference.

Q: Where does the RX 6700S outperform the RX 6700M?

A: The RX 6700S wins in five tests: Passmark DirectX 11 (134 vs 129, -3.7%), Passmark DirectX 9 (195 vs 155, -20.5%), Passmark G2D (739 vs 555, -24.9%), Passmark G3D (15,066 vs 13,536, -10.2%), and Passmark GPU Compute (5,990 vs 5,191, -13.3%).

Q: How do the two compare in Vulkan performance?

A: The RX 6700M leads in Geekbench Vulkan with a score of 90,816 versus 76,150 for the RX 6700S. That is a 19.3% advantage for the RX 6700M.

Q: What is the memory capacity difference?

A: The RX 6700M carries 10 GB of GDDR6 on a 160-bit bus with 320.0 GB/s bandwidth. The RX 6700S has 8 GB of GDDR6 on a 128-bit bus with 224.0 GB/s bandwidth.

Q: Which GPU is more power-efficient based on the data?

A: The RX 6700S has a 80 W TDP, while the RX 6700M has a 135 W TDP. The RX 6700S produces 7.168 TFLOPS FP32 at 80 W, while the RX 6700M produces 11.06 TFLOPS at 135 W. Per watt, the RX 6700S delivers approximately 89.6 GFLOPS/W versus 81.9 GFLOPS/W for the RX 6700M, meaning the RX 6700S is more efficient in raw compute per watt.

Architecture Differences

Both GPUs are built on the RDNA 2.0 architecture and fabricated at TSMC on a 7 nm process. They share the same Radeon RX 6000 series branding and support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The core difference lies in their silicon: the RX 6700M uses the Navi 22 chip, while the RX 6700S uses the Navi 23 chip.

The Navi 22 die in the RX 6700M is substantially larger at 335 mm² and contains 17,200 million transistors, resulting in a transistor density of 51.3 million per mm². The Navi 23 die in the RX 6700S is 237 mm² with 11,060 million transistors, giving a density of 46.7 million per mm². This is a 98 mm² die size difference and a 6,140 million transistor difference.

The RX 6700M has 2,304 shading units, 144 texture mapping units, and 64 ROPs. The RX 6700S reduces this to 1,792 shading units, 112 TMUs, and 64 ROPs. The RX 6700M also has 36 ray tracing cores versus 28 for the RX 6700S.

Neither GPU has tensor cores; both rely on standard RDNA 2 compute units for all processing. The RX 6700M supports PCIe 4.0 x16, while the RX 6700S uses PCIe 4.0 x8, halving the interface width for data transfers.

Memory architecture differs in capacity and width. The RX 6700M uses 10 GB of GDDR6 on a 160-bit bus with effective 16 Gbps memory speed. The RX 6700S uses 8 GB on a 128-bit bus with 14 Gbps effective speed. The bandwidth gap is substantial: 320.0 GB/s versus 224.0 GB/s.

Specification Differences

The two GPUs differ in several key specifications. The RX 6700M has a base clock of 1489 MHz, boost of 2400 MHz, and game clock of 2300 MHz. The RX 6700S has a higher base clock of 1700 MHz but lower boost (2000 MHz) and game clock (1890 MHz). The RX 6700M boosts 400 MHz higher than the RX 6700S.

Memory speed differs: 2000 MHz (16 Gbps effective) for the RX 6700M versus 1750 MHz (14 Gbps effective) for the RX 6700S. The RX 6700M has a 160-bit bus, the RX 6700S a 128-bit bus.

Compute throughput: the RX 6700M reaches 11.06 TFLOPS FP32 and 22.12 TFLOPS FP16 (2:1 ratio). The RX 6700S produces 7.168 TFLOPS FP32 and 14.34 TFLOPS FP16.

Pixel and texture rates reflect the same hierarchy: the RX 6700M outputs 153.6 GPixel/s and 345.6 GTexel/s, while the RX 6700S outputs 128.0 GPixel/s and 224.0 GTexel/s.

The power envelope differs significantly: 135 W for the RX 6700M versus 80 W for the RX 6700S. Both are integrated-on-package designs (IGP) with no power connectors, and their display outputs are portable-device dependent.

Release dates are roughly 7 months apart: the RX 6700M launched May 30, 2021, the RX 6700S on January 3, 2022. Both are marked end-of-life and have no launch MSRP recorded.

Head-to-Head Benchmarks

The RX 6700M dominates the modern API benchmarks. In 3DMark Steel Nomad DX12, it scores 1,845 versus 1,451, a 27.2% lead. Geekbench Metal shows 91,911 versus 72,864, a 26.1% advantage. Geekbench Vulkan gives 90,816 versus 76,150, a 19.3% lead. Geekbench OpenCL sees a smaller but still decisive margin: 77,666 versus 70,965, a 9.4% difference.

The RX 6700M also wins in DirectX 12 (65 vs 58, +12.1%) and DirectX 10 (92 vs 87, +5.7%). These six wins represent the RX 6700M's core strengths: modern graphics APIs and ray tracing workloads.

The RX 6700S wins five tests, and its victories are concentrated in legacy and 2D workloads. The largest margin is in Passmark DirectX 9: 195 versus 155, a 20.5% advantage for the RX 6700S. In Passmark G2D, the RX 6700S scores 739 against 555, a 24.9% lead. The Passmark G3D score favors the RX 6700S at 15,066 versus 13,536, a 10.2% difference. The RX 6700S also leads in GPU compute (5,990 vs 5,191, +13.3%) and DirectX 11 (134 vs 129, +3.7%).

The pattern is clear: the RX 6700M dominates in 12_2-era APIs and modern compute, while the RX 6700S excels in legacy DirectX and general 2D tasks. The RX 6700S wins in DirectX 9 by a wide margin, suggesting its higher base clock (1700 MHz) and lower memory latency may benefit older, less parallel workloads.

The RX 6700M's 27.2% lead in 3DMark Steel Nomad is its biggest victory, while the RX 6700S's 24.9% lead in G2D is its largest. The overall win count is 6-5 in favor of the RX 6700M, but the average score differential (25,633 vs 22,154, a 15.7% gap) reflects the RX 6700M's advantage in more computationally demanding scenarios.

Where Each One Wins

RX 6700M: Modern Graphics and High-Resolution Gaming

The record shows the RX 6700M is superior in DirectX 12 Ultimate workloads, Vulkan, and Metal. Its 27.2% lead in 3DMark Steel Nomad (1,845 vs 1,451) indicates it handles modern game engines better. The 19.3% Vulkan advantage (90,816 vs 76,150) confirms this is not a driver-specific fluke but a hardware-level capability gap. With 10 GB VRAM and 320 GB/s bandwidth, the RX 6700M can manage larger texture sets and higher resolutions without hitting memory constraints. Its 36 ray tracing cores versus 28 provide 28.6% more ray tracing hardware, which aligns with its strength in modern API tests. This is the pick for a user running current games at high settings or using GPU compute in DirectX 12 environments.

RX 6700S: Legacy Titles, 2D, and Power Efficiency

The RX 6700S wins every test in the Passmark suite except DirectX 12 and DirectX 10, including a 20.5% lead in DirectX 9 (195 vs 155) and a 24.9% lead in G2D (739 vs 555). This suggests the higher base clock of 1700 MHz benefits games and applications built on older, less parallel rendering pipelines. The RX 6700S also takes the Passmark G3D (15,066 vs 13,536) and GPU compute (5,990 vs 5,191), indicating it handles general-purpose workloads efficiently despite lower peak FP32. The 80 W power envelope versus 135 W means it is better suited for thinner laptops, longer battery life, and lower thermal loads. For a user playing DirectX 9/11 classics, running 2D productivity apps, or needing compute performance per watt, the RX 6700S delivers wins where the RX 6700M cannot.

The Data Verdict

The RX 6700M is the stronger GPU overall, with a higher percentile (71 vs 67) and a 15.7% average score advantage. It is clearly the main choice for modern gaming and API-heavy tasks. The RX 6700S is a specialized alternative: it wins legacy workloads by margins up to 25%, but these tests represent a shrinking portion of the gaming landscape. The 7-month later release date and lower power draw suggest the RX 6700S targets thin-and-light laptops where the RX 6700M cannot fit. The recommendation is simple: if the workload is modern, take the RX 6700M; if it is legacy or power-constrained, the RX 6700S offers specific wins.