AMD Radeon 760M vs NVIDIA GeForce GTX 980M Comparison

The AMD Radeon 760M and the NVIDIA GeForce GTX 980M represent two very different answers to the same question: how do you get respectable graphics performance into a thin, power-constrained machine? The 760M is a modern integrated GPU built on TSMC's 4 nm process as part of the Phoenix chip, while the GTX 980M is a discontinued MXM mobile module from the Maxwell 2.0 era, manufactured on a 28 nm node. Across the ten recorded head-to-head benchmarks, the GTX 980M wins seven and the 760M wins three, yet the database shows the 760M sitting at the 35th percentile of all GPUs against the 980M's 31st, with an average benchmark score of 6019 versus 5308. The gap between those two pictures tells you most of what you need to know about this matchup.

Where Each One Wins

The GTX 980M owns traditional rasterization gaming. It takes the 3DMark Steel Nomad DirectX 12 test with a score of 649 against 400, a 38.4 percent advantage for the NVIDIA card. It sweeps every PassMark DirectX suite: DirectX 9 by 48 percent (125 versus 65), DirectX 10 by 45.7 percent (35 versus 19), DirectX 11 by 8.8 percent (57 versus 52), and DirectX 12 by 19.4 percent (31 versus 25). The headline PassMark G3D result also goes NVIDIA's way, 7338 to 5310, a 27.6 percent margin. If the workload is older game engines or classic DirectX rendering paths, the 980M is the stronger performer.

The Radeon 760M wins where modern APIs and compute matter. Its biggest victory is Geekbench Vulkan, where it scores 30336 against 17703, a 71.4 percent lead. It also dominates the PassMark G2D test, 890 to 490, an 81.6 percent gap, and edges the PassMark GPU Compute test 2840 to 2816, a narrow 0.9 percent win. OpenCL is the one modern-API test the 980M takes, 23832 to 20255, a 15 percent advantage, so the compute picture is split rather than one-sided. In practical terms: the 760M is the better pick for Vulkan-native workloads and general GPU compute, while the 980M remains the faster pure gaming part in the recorded data.

Architecture Differences

These two chips come from opposite ends of a decade of silicon development. The Radeon 760M is built on TSMC's 4 nm process as the Phoenix implementation of RDNA 3.0, packing 25,390 million transistors into a 178 mm² die, which works out to a density of 142.6M per mm². The GTX 980M's GM204 chip uses the 28 nm node, carries 5,200 million transistors across a much larger 398 mm² die, and lands at just 13.1M per mm². That is roughly an order of magnitude difference in density, and it explains how the 760M fits a full modern GPU into a processor die while the 980M needed a dedicated MXM module.

The resource split is lopsided in the old card's favor on paper. The 980M has 1536 shading units, 96 TMUs, and 64 ROPs, against the 760M's 512 shading units, 32 TMUs, and 16 ROPs. Yet the 760M's clock structure is dramatically more aggressive: it boosts to 2599 MHz from an 800 MHz base, while the 980M runs a 1038 MHz base and 1127 MHz boost. The result is that the smaller modern GPU actually posts the higher theoretical peak figures, at 5.323 TFLOPS FP32 versus 3.462 TFLOPS for the 980M, with the 760M also matching that rate in FP16 at a 1:1 ratio. Texture and pixel throughput still favor NVIDIA, with the 980M rated at 108.2 GTexel/s and 72.13 GPixel/s against 83.17 GTexel/s and 41.58 GPixel/s.

The 760M also carries 8 RT cores, a feature the Maxwell-based 980M simply lacks, and supports DirectX 12 Ultimate (feature level 12_2) versus the 980M's DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4. Memory strategy differs fundamentally: the 980M has dedicated 8 GB of GDDR5 on a 256-bit bus delivering 160.4 GB/s at an effective 5 Gbps, while the 760M shares system memory over its PCIe 4.0 x8 platform, meaning bandwidth and capacity depend entirely on the host system's configuration.

Head-to-Head Benchmarks

Walking through the recorded results in order of magnitude, the 760M's standout number is Geekbench Vulkan at 30336 versus 17703. A 71.4 percent win on a modern cross-platform API is the single largest gap in either direction in this matchup, and it lines up with the 760M's higher FP32 throughput and newer architecture. Its PassMark G2D win, 890 versus 490, is even larger at 81.6 percent, though 2D performance matters less for gaming than for desktop and video workflows.

The 980M's counterpunch is broad rather than deep. Its biggest single margin is PassMark DirectX 9 at 125 versus 65, a 48 percent win, followed closely by DirectX 10 at 35 versus 19 (45.7 percent) and Steel Nomad at 649 versus 400 (38.4 percent). The PassMark G3D aggregate, 7338 versus 5310, confirms the 27.6 percent overall gaming lead, and DirectX 12 goes to NVIDIA 31 to 25. DirectX 11 is the closest gaming test at 57 versus 52, just an 8.8 percent gap, suggesting that in the newest rasterization workloads the two are converging even before the 760M's Vulkan strength comes into play.

Compute splits nearly down the middle. PassMark GPU Compute lands at 2840 for the 760M and 2816 for the 980M, a 0.9 percent difference that is effectively a tie, while Geekbench OpenCL favors the 980M 23832 to 20255 by 15 percent. The composite averages, 6019 for the 760M against 5308 for the 980M, reflect the fact that the 760M's few wins are enormous while the 980M's many wins include several narrow ones.

Specification Differences

The two cards differ on nearly every line of the spec sheet. Process node: 4 nm for the 760M versus 28 nm for the 980M. Architecture: RDNA 3.0 versus Maxwell 2.0. Die size: 178 mm² versus 398 mm². Transistor count: 25,390 million versus 5,200 million. Shading units: 512 versus 1536, TMUs 32 versus 96, ROPs 16 versus 64. Clocks invert that pattern, with the 760M at 800 MHz base and 2599 MHz boost against the 980M's 1038 MHz base and 1127 MHz boost.

Memory is the clearest divide: system-shared on the 760M versus 8 GB of dedicated GDDR5 on a 256-bit bus with 160.4 GB/s of bandwidth on the 980M. Form factor follows from that, with the 760M being an IGP with no power connectors and a 15 W TDP, while the 980M is an MXM module on an MXM-B (3.0) interface. The 760M supports DirectX 12 Ultimate (12_2) and has 8 RT cores; the 980M stops at DirectX 12 (12_1) with no RT hardware. Release dates are far apart, and production status matters too: the 760M is an active product, while the 980M is end-of-life, succeeded by the GeForce 10 Mobile line and preceded by the GeForce 800M series. The 760M's predecessor is the Navi II IGP line.

FAQ

Q: Which GPU is faster for gaming overall?

A: The GTX 980M. It wins the PassMark G3D test 7338 to 5310, a 27.6 percent advantage, and takes five of the six game-oriented benchmarks including 3DMark Steel Nomad at 649 versus 400.

Q: Does the Radeon 760M win anything decisively?

A: Yes. It wins Geekbench Vulkan by 71.4 percent (30336 versus 17703), PassMark G2D by 81.6 percent (890 versus 490), and PassMark GPU Compute by a slim 0.9 percent (2840 versus 2816).

Q: How do they compare in compute workloads?

A: PassMark GPU Compute is effectively tied, 2840 to 2816 in favor of the 760M. Geekbench OpenCL favors the 980M, 23832 to 20255, while Geekbench Vulkan heavily favors the 760M, 30336 to 17703.

Q: How much power does the Radeon 760M use?

A: Its TDP is 15 W, which is what allows it to live inside a processor as an integrated GPU. No TDP figure is recorded for the GTX 980M in the database.

Q: Which GPU has ray tracing support?

A: The 760M, with 8 RT cores and DirectX 12 Ultimate (12_2) support. The 980M has no RT cores and supports DirectX 12 (12_1).

Q: Can you still buy the GTX 980M new?

A: No. It is listed as end-of-life. The 760M is an active product, which matters for driver support and platform availability.

The Verdict

The recorded data points in two directions depending on what you run. For gaming, especially anything built on older DirectX paths, the GTX 980M is the clear choice: 27.6 percent ahead in PassMark G3D, 38.4 percent ahead in Steel Nomad, and ahead in every DirectX suite from 9 through 12. Its dedicated 8 GB of GDDR5 with 160.4 GB/s of bandwidth also removes the memory variability that comes with a system-shared design.

For modern API work and efficiency-driven builds, the 760M makes its case. The 71.4 percent Vulkan win is the largest margin on the board, its 15 W TDP and integrated form factor mean no module, no dedicated cooling arrangement, and no shared-bandwidth compromise on capacity, and its higher aggregate average score (6019 versus 5308) and better percentile placement (35 versus 31) show the wins it does get are substantial. It also carries RT cores and DirectX 12 Ultimate support the 980M lacks, and it is an active product while the 980M is discontinued. Builders prioritizing classic game performance should follow the seven benchmark wins toward the GTX 980M; anyone working with Vulkan or compute, or building around a modern low-power platform, gets the stronger long-term package in the Radeon 760M.