AMD Radeon 890M vs NVIDIA GeForce GTX 960 Comparison

The NVIDIA GeForce GTX 960 and AMD Radeon 890M occupy the same performance percentile (45th) in the database, yet their benchmark profiles could hardly be more different. The GTX 960 is a 2015 discrete desktop card built on a 28 nm process, while the 890M is a 2024 integrated graphics processor (IGP) on a 4 nm node with system-shared memory. Their average benchmark scores are nearly identical—9273 for the GTX 960 versus 9210 for the 890M—but this parity masks a decisive generational shift in where and how performance is delivered.

Head-to-Head Benchmarks

The head-to-head results are one-sided in favor of the AMD Radeon 890M, which wins all three shared tests. The largest margin appears in Geekbench Vulkan, where the 890M scores 40808 against the GTX 960's 9231, a delta of -77.4% from the AMD part's perspective (meaning the GTX 960 trails by that amount). This is not a marginal lead; the 890M delivers more than quadruple the Vulkan throughput. In 3DMark Steel Nomad DX12, the 890M posts 590 points versus 162 for the GTX 960, a -72.5% delta, again a roughly 3.6x advantage for the integrated part. The closest contest is Geekbench OpenCL, where the 890M scores 37254 and the GTX 960 scores 18925, a -49.2% delta; even here, the AMD IGP is nearly twice as fast.

These results indicate that the 890M's architectural advantages—RDNA 3.5 with 16 ray accelerators and a 2.9 GHz boost clock—translate into massive gains in modern, low-level APIs. The GTX 960's Maxwell 2.0 architecture, with a 1178 MHz boost, cannot keep pace. Notably, the 890M also has a PassMark G3D score of 8076 and a PassMark GPU Compute score of 4157, while the GTX 960 lacks any PassMark entries in the data, suggesting the older card is not tested in those workloads. The 890M's PassMark DirectX 9 and DirectX 11 scores of 97 and 73 respectively are far lower than its G3D score, indicating that its strength lies in compute and modern APIs rather than legacy rasterization.

The GTX 960's only comparative bright spot is its Geekbench Metal score of 8773, a test the 890M does not appear in. However, Metal is an Apple-specific API, and the 890M's absence from that test likely reflects platform targeting rather than a performance deficiency. In the shared tests, the 890M's victory is absolute, with a 3-0 win count in the head-to-head table.

The Verdict

From the data, the AMD Radeon 890M is the superior performer in every directly comparable benchmark. Its average benchmark score of 9210 is only 0.7% below the GTX 960's 9273, but this aggregate figure is misleading because the 890M's scores are dragged down by its PassMark DirectX 10 (33), DirectX 11 (73), and DirectX 12 (37) results, which are legacy 2D-oriented tests. In compute-heavy workloads—OpenCL, Vulkan, and 3DMark Steel Nomad—the 890M is 49% to 77% faster. The GTX 960's nearest rival list shows it trading blows with the GTX 465 (9294, -0.2%) and GTX 850M (9302, -0.3%), while the 890M's list places the GTX 960 as its second-closest rival (-0.7%), behind only the AMD Radeon Vega 8 (-0.1%).

Who should pick which? Strictly from the numbers, the 890M is the choice for any workload using Vulkan, OpenCL, or DX12, where it delivers 2x to 4x the performance. The GTX 960 has no winning benchmark category in the head-to-head set, so its only argument is its 2 GB of dedicated GDDR5 memory versus the 890M's system-shared memory. However, the 890M's memory bandwidth is listed as "System Dependent," meaning its effective throughput varies with the host system's RAM, while the GTX 960 has a fixed 112.2 GB/s. For users who need a discrete card with a dedicated 128-bit memory bus, the GTX 960 retains a structural advantage, but for raw compute and modern API performance, the 890M is decisively ahead. The 890M also does this at a 15 W TDP versus the GTX 960's 120 W, a factor that cannot be ignored for mobile or low-power systems.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The NVIDIA GeForce GTX 960 has an average benchmark score of 9273, while the AMD Radeon 890M scores 9210. The difference is 63 points, or about 0.7% in favor of the GTX 960.

Q: In which benchmark does the AMD Radeon 890M show its largest lead?

A: The largest lead is in Geekbench Vulkan, where the 890M scores 40808 versus the GTX 960's 9231, a delta of -77.4% (the GTX 960 trails by that margin).

Q: Does the NVIDIA GeForce GTX 960 win any shared benchmark?

A: No. In the three head-to-head benchmarks (3DMark Steel Nomad DX12, Geekbench OpenCL, Geekbench Vulkan), the AMD Radeon 890M wins all three, with the GTX 960 trailing by 72.5%, 49.2%, and 77.4% respectively.

Q: What is the memory configuration difference between the two?

A: The GTX 960 has 2 GB of dedicated GDDR5 memory on a 128-bit bus with a fixed 112.2 GB/s bandwidth. The 890M uses System Shared memory with a System Shared bus width and System Dependent bandwidth, meaning it relies on the host system's RAM.

Q: How do their process nodes compare?

A: The GTX 960 is built on a 28 nm process at TSMC, while the 890M is on a 4 nm process, also at TSMC. The 890M has a much higher transistor density at 145.9M / mm² versus 12.9M / mm² for the GTX 960.

Q: Which GPU has ray tracing cores?

A: Only the AMD Radeon 890M has ray tracing cores, with 16 of them. The GTX 960 has none, as its Maxwell 2.0 architecture predates hardware ray tracing.

Specification Differences

The two GPUs differ in nearly every fundamental specification. The GTX 960 uses the GM206 chip with a 28 nm process, while the 890M uses the Strix Point chip on a 4 nm node. Transistor counts diverge sharply: the GTX 960 has 2,940 million transistors on a 228 mm² die, while the 890M packs 34,000 million on a 233 mm² die. This yields a transistor density of 12.9M / mm² for NVIDIA versus 145.9M / mm² for AMD.

Clock speeds are inverted in nature: the GTX 960 has a base clock of 1127 MHz and boost of 1178 MHz, while the 890M has a low 400 MHz base but a 2900 MHz boost. Memory is another stark contrast—the GTX 960 has 2 GB of GDDR5 with a 128-bit bus and 112.2 GB/s bandwidth, whereas the 890M uses system-shared memory with system-dependent bandwidth. The TDP gap is enormous: 120 W for the GTX 960 versus 15 W for the 890M. The GTX 960 requires a 1x 6-pin power connector and a 300 W suggested PSU, while the 890M is an IGP with no power connectors and no suggested PSU. The bus interface differs: PCIe 3.0 x16 for the GTX 960 versus PCIe 4.0 x8 for the 890M. Display outputs also differ, with the GTX 960 offering 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2, while the 890M's outputs are portable-device dependent.

Architecture Differences

Architecturally, these are two completely different generations. The GTX 960 is built on Maxwell 2.0, part of the GeForce 900 generation, with a release date of 2015-01-21 and end-of-life status. The 890M is RDNA 3.5, part of the Navi III IGP (Strix Point Mobile) generation, released on 2024-07-14 and still active. The GTX 960 has 1024 shading units, 64 TMUs, and 32 ROPs; the 890M has the same counts of shading units, TMUs, and ROPs, but adds 16 ray tracing cores. The 890M also supports DirectX 12 Ultimate (12_2), while the GTX 960 only reaches DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4.

Fill rates show the 890M's advantage: it has a pixel rate of 92.80 GPixel/s and texture rate of 185.6 GTexel/s, versus 37.70 GPixel/s and 75.39 GTexel/s for the GTX 960. FP32 compute is also one-sided: the 890M delivers 5.939 TFLOPS, more than double the GTX 960's 2.413 TFLOPS. The 890M also supports FP16 at a 1:1 ratio (5.939 TFLOPS), while the GTX 960 has no listed FP16 capability. The 890M's predecessor is the Navi II IGP, and it has no successor, while the GTX 960's predecessor is GeForce 700 and its successor is GeForce 10. The GTX 960 has a launch MSRP of 199 USD, while the 890M has no launch MSRP listed. The 890M's 16 ray accelerators are a feature class the GTX 960 cannot offer, and the 1:1 FP16 rate suggests a compute-oriented design that the Maxwell part lacks.