NVIDIA GeForce GTX 950M vs NVIDIA GeForce GTX 960M Comparison

The NVIDIA GeForce GTX 960M and NVIDIA GeForce GTX 950M are siblings in the deepest sense: both ship on the same GM107 die, both belong to the Maxwell-based GeForce 900M mobile generation, and both carry an identical count of shading units, texture units, and render outputs. Yet the recorded benchmark data shows the GTX 960M winning every head-to-head test, with margins ranging from double digits in compute to a decisive gap in the newer Vulkan API. The difference comes down almost entirely to memory and clock configuration, and the database makes that divergence easy to trace.

Head-to-Head Benchmarks

Across the two recorded tests, the GTX 960M swept the matchup with two wins against zero for the GTX 950M, and the margins were substantial in both cases.

The Geekbench OpenCL result sets the tone: the GTX 960M scored 11045 against 9745 for the GTX 950M, a 13.3 percent advantage in general GPU compute. That is a meaningful gap between two parts built on the same silicon, and it reflects both the GTX 960M's higher clock speeds and, critically, its much faster memory subsystem.

The Geekbench Vulkan test widens the gap dramatically. Here the GTX 960M posted 8245 against 6525 for the GTX 950M, a 26.4 percent lead. Vulkan workloads tend to expose memory bandwidth limitations more harshly than older APIs, and the numbers bear this out: the GTX 950M's DDR3 memory delivers only 28.80 GB/s of bandwidth, while the GTX 960M's GDDR5 pushes 80.19 GB/s, nearly three times the throughput on the same 128-bit bus. When a workload stresses sustained data movement, the DDR3 configuration becomes the bottleneck and the Vulkan delta roughly doubles compared to the OpenCL result.

Aggregated, the GTX 960M averages 9645 across the recorded benchmarks versus 8135 for the GTX 950M. The database places the GTX 960M at the 46th percentile against all GPUs, with the GTX 950M four points behind at the 42nd percentile. Both are mid-table performers in the historical record, but the GTX 960M sits meaningfully higher.

The rivals clustered around each card confirm the tier separation. The GTX 960M's average score of 9645 lands within a fraction of a percent of the NVIDIA Quadro K5000 at 9637 (0.1 percent ahead), the AMD Radeon Pro WX 2100 at 9653 (0.1 percent behind), the NVIDIA Quadro P4000 at 9665 (0.2 percent behind), and the NVIDIA Tesla C2070 at 9716 (0.7 percent behind). The GTX 950M, meanwhile, clusters with the AMD Radeon R9 M360 at 8129 (0.1 percent ahead), the NVIDIA GeForce GTX 980 at 8167 (0.4 percent behind), the NVIDIA GeForce 945M at 8099 (0.4 percent ahead), and the NVIDIA GRID K2 at 8080 (0.7 percent ahead). The two cards never overlap in their rival sets, reinforcing that the database treats them as occupants of adjacent but distinct performance tiers.

Where Each One Wins

On the recorded evidence, the GTX 950M wins nothing. It lost both benchmarks, by 13.3 percent in OpenCL and 26.4 percent in Vulkan, and its aggregate average trails by a wide margin. The head-to-head record is unambiguous: two tests, two GTX 960M victories.

The reason is structural rather than architectural. The two cards share the same GM107 chip, the same 640 shading units, 40 texture units, and 16 render outputs, and the same 4 GB memory capacity on a 128-bit bus. What separates them is implementation. The GTX 960M runs a base clock of 1097 MHz with a 1176 MHz boost, against 993 MHz base and 1124 MHz boost for the GTX 950M. Note that the boost clocks are close, which explains why the raw compute figures are relatively near each other: the GTX 960M manages 1.505 TFLOPS of FP32 against 1438.7 GFLOPS for the GTX 950M, a difference of under 5 percent.

The real gulf is memory. The GTX 960M uses GDDR5 running at 1253 MHz, 5 Gbps effective, for 80.19 GB/s of bandwidth. The GTX 950M uses DDR3 at 900 MHz, 1800 Mbps effective, for just 28.80 GB/s. Any workload that depends on bandwidth rather than pure ALU throughput, and the Vulkan result strongly suggests modern API workloads fall into that category, favors the GTX 960M disproportionately. The pixel and texture rates echo this: the GTX 960M reaches 18.82 GPixel/s and 47.04 GTexel/s against 17.98 GPixel/s and 44.96 GTexel/s for the GTX 950M, gaps consistent with the clock differences but far smaller than the benchmark gaps.

The practical reading: for gaming and modern API workloads, the GTX 960M is the clear choice from the data. For tasks that are purely compute-bound and lightly dependent on memory traffic, the gap narrows considerably, though it never closes.

The Verdict

The data supports one recommendation: between these two cards, take the GTX 960M. It wins every recorded benchmark, holds a four-point percentile advantage across the full GPU database, and clusters with stronger rivals. The GTX 950M's only recorded advantages are incidental: it is implemented as an integrated GPU part on a PCIe 3.0 x8 interface rather than an MXM module, which may matter for the type of machine it inhabits but says nothing about performance.

Both cards are end-of-life, both launched on the same date in March of 2015, and both draw 75 W with no external power connectors, so there is no efficiency argument to be made in either direction from the recorded data. With identical API support, including DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4, and identical 4 GB frame buffers, the deciding factors reduce purely to clock speeds and memory type, and the GTX 960M wins on both counts. Anyone choosing between the two for gaming, compute, or general GPU work should pick the GTX 960M on the strength of a clean sweep in the benchmark record.

FAQ

Q: Which GPU is faster overall, the GTX 960M or the GTX 950M?

A: The GTX 960M. It won both recorded benchmarks, leads the average score 9645 to 8135, and ranks at the 46th percentile of all GPUs versus the GTX 950M's 42nd percentile.

Q: How large is the performance gap?

A: The GTX 960M leads by 13.3 percent in Geekbench OpenCL and 26.4 percent in Geekbench Vulkan, meaning the gap grows in the more modern API test.

Q: Do the two GPUs use the same chip?

A: Yes. Both are built on the GM107 die of Maxwell architecture, manufactured by TSMC on a 28 nm process, with 1,870 million transistors on a 148 mm² die at a density of 12.6M transistors per mm².

Q: Why is the GTX 960M faster if the chips are identical?

A: Two reasons. It clocks higher, with a 1176 MHz boost versus 1124 MHz, and it uses GDDR5 memory delivering 80.19 GB/s of bandwidth versus the GTX 950M's DDR3 at 28.80 GB/s. The memory difference is the dominant factor.

Q: How do they compare on raw compute?

A: The FP32 figures are close: 1.505 TFLOPS for the GTX 960M versus 1438.7 GFLOPS for the GTX 950M, a gap of under 5 percent. The benchmark gaps are much larger because real workloads depend heavily on memory bandwidth.

Q: Are both cards still in production?

A: No. Both are listed as end-of-life, having launched on the same day in March 2015, with the GeForce 800M series as predecessor and GeForce 10 Mobile as successor.

Architecture Differences

In most architectural respects these cards are the same design. Both use NVIDIA's Maxwell architecture in the GM107 configuration: 640 shading units, 40 texture units, and 16 render outputs, fabricated at TSMC on the 28 nm node with 1,870 million transistors packed into a 148 mm² die. Neither part carries RT cores or tensor cores, as expected for this generation. Both support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.4, and both carry 4 GB of memory on a 128-bit bus.

The differences are in execution. Clocks diverge modestly: the GTX 960M runs a 1097 MHz base and 1176 MHz boost, while the GTX 950M runs 993 MHz base and 1124 MHz boost. The resulting throughput figures are correspondingly close, with pixel rates of 18.82 versus 17.98 GPixel/s and texture rates of 47.04 versus 44.96 GTexel/s.

The memory subsystem is where the designs part ways. The GTX 960M pairs GDDR5 running at 1253 MHz (5 Gbps effective) for 80.19 GB/s of bandwidth. The GTX 950M pairs DDR3 running at 900 MHz (1800 Mbps effective) for just 28.80 GB/s. On identical bus widths and capacities, that memory choice alone accounts for the bulk of the benchmark divergence.

Platform implementation differs too. The GTX 960M is an MXM module on an MXM-B (3.0) interface, while the GTX 950M is an integrated GPU part on PCIe 3.0 x8. Both are rated at 75 W TDP with no power connectors, both have portable-device-dependent display outputs, and both share the same launch date and lifecycle status. Same silicon, same architecture, same feature set: the GTX 960M simply executes the design with faster memory and higher clocks, and the benchmark record rewards it with a clean sweep.