NVIDIA GeForce GTX 780M vs NVIDIA GeForce GTX 950 Comparison

The Verdict

The recorded data separates these two NVIDIA parts by their workload strengths, not by overall supremacy. The GTX 950 takes the aggregate benchmark average with 13,189 points against the GTX 780M’s 11,261 points, a lead of roughly 17%. That overall margin comes almost entirely from modern API tests. In Geekbench OpenCL, the GTX 950 scores 15,662 versus 12,769 for the 780M, a 22.7% advantage. In Geekbench Vulkan, the gap widens to 31.8% (16,734 versus 12,696). The 780M fights back in Geekbench Metal, where it scores 8,319 against 7,172, a 13.8% win. The verdict is straightforward: the GTX 950 is the stronger choice for OpenCL and Vulkan workloads, while the 780M retains an edge specifically in Metal-based applications. Desktop builders should favor the GTX 950 for its modern feature set and API support. Laptop or MXM users with a 780M already installed should not expect a generational leap if they switch, except in Vulkan scenarios where the 950 is decisively faster.

Architecture Differences

The two GPUs come from different NVIDIA architectures and different eras. The GTX 950 uses the GM206 chip on Maxwell 2.0, part of the GeForce 900 generation, fabricated on a 28 nm process at TSMC. The GTX 780M uses the GK104 chip on Kepler, part of the GeForce 700M generation, also on a 28 nm TSMC process. Both share the same process node, but the transistor counts differ notably. The 780M packs 3,540 million transistors on a 294 mm² die, while the 950 has 2,940 million transistors on a smaller 228 mm² die. The density numbers are close: 12.9M per mm² for the 950 versus 12.0M per mm² for the 780M, so the 950 achieves slightly better packing efficiency.

The compute resources tell a different story. The 780M carries 1,536 shading units, 128 texture units, and 32 ROPs. The 950 has 768 shading units, 48 texture units, and 32 ROPs. Despite having half the shader count, the 950 posts higher pixel rate (38.02 GPixel/s versus 25.50 GPixel/s) because its boost clock reaches 1188 MHz against the 780M’s 797 MHz. The 950’s texture rate is lower at 57.02 GTexel/s versus 102.0 GTexel/s, reflecting its fewer TMUs. In raw FP32 throughput, the 780M claims 2.448 TFLOPS versus the 950’s 1.825 TFLOPS, but that theoretical advantage does not translate to benchmark wins in the recorded tests.

Memory configurations also diverge. The 950 has 2 GB of GDDR5 on a 128-bit bus, yielding 105.8 GB/s bandwidth. The 780M has 4 GB of GDDR5 on a 256-bit bus, providing 160.0 GB/s. The 780M’s larger frame buffer and wider bus favor high-resolution textures and anti-aliasing, while the 950’s smaller pool may limit very demanding scenes. The 950’s memory clock is higher at 1653 MHz (6.6 Gbps effective) versus 1250 MHz (5 Gbps effective) for the 780M. Power and interface differ as well: the 950 is a dual-slot desktop card rated at 90 W TDP with a 1x 6-pin connector and PCIe 3.0 x16 interface, while the 780M is an MXM module rated at 122 W with no power connectors and an MXM-B (3.0) interface. API support is a key differentiator. The 950 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The 780M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The 950’s newer feature level explains part of its Vulkan advantage.

Head-to-Head Benchmarks

The head-to-head results show a clear split by API. In Geekbench Metal, the GTX 780M wins with 8,319 points against 7,172 for the GTX 950. The delta is 13.8% in favor of the 780M. This is the only test where the older Kepler part comes out ahead. The Metal result is notable because the 780M’s higher shader count and wider memory bus may compensate for its lower clocks in that specific workload. Meanwhile, the GTX 950 dominates the other two tests by larger margins. In Geekbench OpenCL, the 950 scores 15,662 versus 12,769, a 22.7% lead. In Geekbench Vulkan, the 950 scores 16,734 versus 12,696, a 31.8% lead. These are substantial gaps, not marginal differences. The Vulkan result is particularly striking: the 950 outperforms the 780M by nearly a third in that API. The pattern suggests that the 950’s Maxwell 2.0 architecture and newer driver path handle compute and modern graphics APIs more efficiently, despite having fewer shading units and lower theoretical FP32.

The overall win count is 2 for the GTX 950 and 1 for the GTX 780M. The average benchmark score reinforces the 950’s position: 13,189 versus 11,261, a difference of 1,928 points. The 950 sits at percentile 53 among all GPUs, while the 780M sits at percentile 50. The nearest rivals in the database place the 950 close to the NVIDIA GeForce GTX 480 (13,300 average, 0.8% higher), the AMD Radeon Pro 555X (13,321 average, 1% higher), and the AMD FirePro M6100 (13,354 average, 1.2% higher). The 780M’s nearest rivals include the AMD Radeon Pro WX 3200 (11,228 average, 0.3% lower) and the AMD FirePro W4300 (11,225 average, 0.3% lower), placing it in a lower performance tier overall.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The GTX 950 leads with an average benchmark score of 13,189 against the GTX 780M’s 11,261. The 950 also holds a higher percentile ranking at 53 versus 50.

Q: In which test does the GTX 780M beat the GTX 950?

A: The 780M wins only in Geekbench Metal, scoring 8,319 versus 7,172, a 13.8% advantage. The 950 wins the OpenCL and Vulkan tests by 22.7% and 31.8% respectively.

Q: How do their memory configurations compare?

A: The 780M has 4 GB of GDDR5 on a 256-bit bus with 160.0 GB/s bandwidth. The 950 has 2 GB of GDDR5 on a 128-bit bus with 105.8 GB/s bandwidth. The 780M’s memory clock is lower at 1250 MHz versus 1653 MHz for the 950.

Q: What API level does each GPU support for DirectX?

A: The GTX 950 supports DirectX 12 (12_1), while the GTX 780M supports DirectX 12 (11_0). Both support OpenGL 4.6, but the 950’s Vulkan version is 1.4 versus 1.2.175 for the 780M.

Q: Which GPU has more shading units and higher theoretical FP32?

A: The GTX 780M has 1,536 shading units and 2.448 TFLOPS FP32. The GTX 950 has 768 shading units and 1.825 TFLOPS. Despite this, the 950 wins two of three benchmarks, indicating architectural efficiency matters more than raw counts.

Q: What are the power and interface differences?

A: The 950 is a dual-slot desktop card rated at 90 W TDP with a 1x 6-pin power connector and PCIe 3.0 x16 interface. The 780M is an MXM module rated at 122 W with no power connectors and an MXM-B (3.0) interface.

Where Each One Wins

The GTX 950 is the clear winner for compute-heavy and modern API workloads. Its OpenCL score of 15,662 and Vulkan score of 16,734 demonstrate strong performance in general-purpose GPU computing and Vulkan-based games or applications. The 22.7% and 31.8% margins over the 780M are not subtle. The 950’s DirectX 12 (12_1) feature level and Vulkan 1.4 support also position it better for current and future titles that lean on these APIs. Its lower TDP of 90 W makes it a reasonable fit for standard desktop builds, and its 202 mm length fits most mid-tower cases. The 950’s nearest rivals hover around its average score (GTX 480 at 13,300, Radeon Pro 555X at 13,321), so it sits in a familiar performance band for mid-range desktop GPUs.

The GTX 780M wins specifically in Metal-based workloads. Its Geekbench Metal score of 8,319 beats the 950 by 13.8%. For users running macOS or Metal-accelerated applications, the 780M has a measurable advantage. Its 4 GB frame buffer and 256-bit bus provide 160.0 GB/s bandwidth, which helps in texture-heavy scenes and higher resolutions. The 780M also has more shading units (1,536 versus 768) and higher theoretical FP32 (2.448 TFLOPS), which may benefit workloads that scale with raw shader throughput rather than architecture efficiency. However, its 122 W TDP and MXM form factor limit it to laptops or specialized systems. The 780M’s nearest rivals, such as the Radeon Pro WX 3200 (11,228 average) and FirePro W4300 (11,225 average), indicate it performs below the 950’s tier in aggregate.

For a desktop user prioritizing OpenCL or Vulkan performance, the GTX 950 is the only rational pick. For a laptop user with a 780M who needs Metal performance, the 780M remains viable. The two GPUs do not compete in the same physical space, so the choice largely depends on platform and workload. The data does not support a blanket recommendation for one over the other; it supports a recommendation based on the API in question. In Vulkan, the 950 is 31.8% faster, which is a decisive margin. In Metal, the 780M is 13.8% faster, which is meaningful but smaller. The 950’s two wins out of three, plus its higher average score and percentile, make it the stronger overall product in the database’s recorded measurements.