AMD Radeon RX 6700M vs NVIDIA GeForce GTX 980 Ti Comparison

FAQ

Q: Which GPU wins in the recorded benchmark suite?

A: The AMD Radeon RX 6700M wins all four head-to-head tests in the database. It takes the 3DMark Steel Nomad DX12 test, Geekbench Metal, Geekbench OpenCL, and Geekbench Vulkan. The NVIDIA GeForce GTX 980 Ti records zero wins in direct comparisons.

Q: How large is the performance gap in the DX12 test?

A: In 3DMark Steel Nomad DX12, the RX 6700M scores 1845 against the GTX 980 Ti's 1321, a delta of -28.4% from the NVIDIA card's perspective. That places the AMD part roughly 28% ahead of the older Maxwell flagship in this specific workload.

Q: What is the most dramatic single-test difference?

A: Geekbench Metal shows the widest margin. The RX 6700M posts 91911 while the GTX 980 Ti manages 19520, a delta of -78.8%. This is a near 4.7x gap in favor of the AMD mobile GPU, though Metal is a workload where the newer architecture clearly excels.

Q: How do the two GPUs compare in overall average score?

A: The GTX 980 Ti has an average benchmark score of 28020 across all recorded tests, placing it in the 73rd percentile of all GPUs. The RX 6700M averages 25633, sitting in the 71st percentile. Despite losing every head-to-head test, the GTX 980 Ti actually holds a slightly higher overall average due to its broader benchmark set.

Q: What are the closest rivals to each card according to the database?

A: For the GTX 980 Ti, the nearest rival is the AMD Radeon Pro W5500X with a delta of +0.2%, followed by the AMD FirePro S7150 at -0.3%. For the RX 6700M, the AMD Radeon Pro W5700 is closest at -0.4%, with the NVIDIA GeForce RTX 3080 Ti Mobile also at -0.4%.

Q: Which card has the higher transistor density?

A: The RX 6700M is built on TSMC's 7 nm process and packs 17,200 million transistors into a 335 mm² die, yielding 51.3M transistors per mm². The GTX 980 Ti uses TSMC's 28 nm node with 8,000 million transistors on a 601 mm² die, which works out to 13.3M per mm². The AMD chip is roughly 3.9x denser.

The Verdict

The data points to a clear generational split. For users prioritizing modern API workloads, ray tracing capable hardware, and mobile form factors, the AMD Radeon RX 6700M is the obvious choice. It wins every recorded head-to-head benchmark, often by substantial margins, and brings features the NVIDIA card simply does not have, such as dedicated ray accelerators and DirectX 12 Ultimate support.

However, the overall average benchmark score tells a more nuanced story. The GTX 980 Ti records 28020 versus the RX 6700M's 25633, a difference of roughly 9%. This is partly because the GTX 980 Ti has a smaller benchmark footprint in the database, and its scores come from tests where it performs closer to its peers. The RX 6700M's percentile rank of 71 versus the GTX 980 Ti's 73 reinforces this: despite winning the direct comparisons, the AMD part sits slightly lower relative to the entire GPU population.

For a desktop user with legacy PCIe 3.0 infrastructure and a need for multiple display outputs, the GTX 980 Ti remains a functional, if dated, option. It offers 6 GB of GDDR5 on a 384-bit bus, dual-slot cooling, and standard display connectivity. The RX 6700M, by contrast, is an integrated mobile part with no dedicated power connectors, no standalone dimensions, and portable-device-dependent outputs. It belongs in a laptop, not a desktop tower.

The verdict hinges on context. In a laptop, the RX 6700M is the only realistic choice of these two. In a desktop, the GTX 980 Ti has the physical and connectivity profile that fits traditional builds, but it surrenders significant performance in modern DX12 and compute workloads. Buyers should match the GPU to the platform.

Head-to-Head Benchmarks

The RX 6700M dominates every recorded direct comparison, but the margins vary widely by workload. In 3DMark Steel Nomad DX12, the AMD card scores 1845 against 1321 for the GTX 980 Ti, a delta of -28.4%. This is a meaningful lead in a modern DirectX 12 test, reflecting the RDNA 2 architecture's ability to handle current rendering features more efficiently than Maxwell 2.0.

The Geekbench Metal test produces the most extreme result. The RX 6700M posts 91911, while the GTX 980 Ti manages only 19520, a delta of -78.8%. This is not a close contest by any measure. Metal is Apple's graphics API, and the RX 6700M's score suggests far better optimization for that compute stack, likely aided by its FP16 throughput of 22.12 TFLOPS (2:1), which the GTX 980 Ti does not expose.

Geekbench OpenCL narrows the gap somewhat but still favors AMD. The RX 6700M scores 77666 against 43513, a delta of -44%. OpenCL is a cross-vendor compute standard, and while the Maxwell GPU remains competitive in raw FP32 (6.060 TFLOPS versus 11.06 TFLOPS), the AMD card's higher shading unit count and newer architecture carry it to a decisive win.

Geekbench Vulkan follows a similar pattern. The RX 6700M records 90816 versus 47724, a delta of -47.4%. Vulkan is the low-level API that benefits from efficient driver overhead and hardware scheduling, areas where RDNA 2.0 has clear advantages over a 2015-era design. The GTX 980 Ti's Vulkan support is listed at 1.4, matching the AMD card, but the underlying hardware efficiency is not comparable.

Across all four tests, the RX 6700M wins with deltas ranging from -28.4% to -78.8%. The smallest margin is in DX12, which is also the most modern rendering path. This suggests that as workloads shift toward DX12 and Vulkan, the AMD part extends its lead, while the GTX 980 Ti's best hope lies in older or less compute-intensive scenarios, none of which appear in the head-to-head set.

Specification Differences

The two GPUs differ on nearly every fundamental specification. Memory capacity favors AMD: 10 GB of GDDR6 versus 6 GB of GDDR5. The bus width, however, favors NVIDIA: 384-bit versus 160-bit. The result is that memory bandwidth is close, with the GTX 980 Ti at 336.6 GB/s and the RX 6700M at 320.0 GB/s. The AMD card uses faster memory at 16 Gbps effective versus 7 Gbps effective, but the narrower bus limits the total throughput.

Clock speeds show the process node advantage. The RX 6700M boosts to 2400 MHz with a game clock of 2300 MHz, while the GTX 980 Ti boosts to 1076 MHz with a base of 1000 MHz. The AMD part runs more than twice as fast at boost. Shading units favor NVIDIA at 2816 versus 2304, but texture mapping units and ROPs are split: the GTX 980 Ti has 176 TMUs and 96 ROPs, while the RX 6700M has 144 TMUs and 64 ROPs. Pixel rate goes to AMD at 153.6 GPixel/s versus 103.3 GPixel/s, and texture rate is similarly lopsided at 345.6 GTexel/s versus 189.4 GTexel/s.

FP32 compute strongly favors the RX 6700M at 11.06 TFLOPS against 6.060 TFLOPS. The AMD card also lists FP16 at 22.12 TFLOPS with a 2:1 ratio, a feature the GTX 980 Ti does not specify. Power draw is drastically different: 135 W for the RX 6700M versus 250 W for the GTX 980 Ti. The NVIDIA card requires a 600 W suggested PSU and 1x 6-pin plus 1x 8-pin power connectors, while the AMD part uses no connectors and has no suggested PSU, consistent with its integrated mobile design.

Bus interface differs as well: PCIe 4.0 x16 for the RX 6700M versus PCIe 3.0 x16 for the GTX 980 Ti. Display outputs are another major split. The GTX 980 Ti offers 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2. The RX 6700M is listed as "Portable Device Dependent," meaning it has no fixed output configuration. Physical dimensions are present for the NVIDIA card (267 mm length, 111 mm height, 40 mm width) but absent for the AMD part. The GTX 980 Ti is dual-slot; the RX 6700M is an IGP.

Architecture Differences

The GTX 980 Ti uses the GM200 chip on NVIDIA's Maxwell 2.0 architecture, fabricated on TSMC's 28 nm node. It belongs to the GeForce 900 generation and carries 8,000 million transistors on a 601 mm² die. This is a large, power-hungry desktop design from 2015, now end-of-life. Its predecessor was the GeForce 700 series, and its successor is the GeForce 10 series. It supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. There are no dedicated ray tracing or tensor cores.

The RX 6700M uses the Navi 22 chip on AMD's RDNA 2.0 architecture, built on TSMC's 7 nm node. It is part of the Radeon RX 6000 series and the Navi Mobile (RX 6000M) generation, released in 2021. The die is 335 mm² with 17,200 million transistors, giving it a far higher transistor density of 51.3M per mm² versus 13.3M per mm² for the NVIDIA part. Its predecessor is Polaris Mobile, and it has no listed successor. It supports DirectX 12 Ultimate (12_2), which includes features beyond the 12_1 level of the GTX 980 Ti, plus OpenGL 4.6 and Vulkan 1.4.

The most significant architectural difference is the presence of 36 ray tracing cores in the RX 6700M. The GTX 980 Ti has none. This enables hardware-accelerated ray tracing on the AMD part, a capability absent from the Maxwell design. The RX 6700M also supports FP16 compute at a 2:1 ratio, doubling its FP32 throughput for compatible workloads. The GTX 980 Ti does not list FP16 performance.

The process node gap is stark: 7 nm versus 28 nm. This drives the RX 6700M's higher clocks, lower power draw, and greater transistor density. The AMD card achieves 135 W TDP despite having more than twice the transistor count, thanks to the denser, more efficient manufacturing process. The GTX 980 Ti's 250 W TDP reflects its older, larger node and desktop-oriented design.

Memory architecture also differs fundamentally. The GTX 980 Ti uses GDDR5 on a 384-bit bus, while the RX 6700M uses GDDR6 on a 160-bit bus. The AMD card compensates for the narrower bus with faster memory clocks and a larger 10 GB capacity. The NVIDIA card leans on its wider bus for bandwidth, though the final numbers are close. The RX 6700M's PCIe 4.0 interface doubles the theoretical link bandwidth of the GTX 980 Ti's PCIe 3.0, which matters for systems with newer CPUs and storage.