NVIDIA GeForce GTX 680M vs NVIDIA GeForce GTX 950M Comparison

The Verdict

The recorded benchmark data positions the NVIDIA GeForce GTX 950M as the stronger overall mobile graphics solution. In the single shared benchmark, Geekbench OpenCL, the GTX 950M scores 9,745 against the GTX 680M’s 9,230, a 5.6% advantage. This places the GTX 950M at the 42nd percentile of all GPUs in the database, while the GTX 680M sits at the 39th percentile. The average benchmark score reinforces this ordering: the GTX 950M averages 8,135, while the GTX 680M averages 7,023.

The GTX 950M’s nearest rivals in the database include the AMD Radeon R9 M360 at 8,129 (a 0.1% difference), the NVIDIA GeForce GTX 980 at 8,167 (0.4% behind), and the NVIDIA GeForce 945M at 8,099 (0.4% ahead). The GTX 680M, by contrast, sits near the NVIDIA T600 at 7,035 (0.2% behind) and the NVIDIA GeForce GTX 675M at 6,946 (1.1% ahead). These adjacent scores show that the GTX 950M competes in a higher performance tier than the GTX 680M, despite the older card’s larger chip and wider memory bus.

For users choosing between these two end-of-life mobile GPUs, the GTX 950M is the better pick for general compute workloads and OpenCL-accelerated tasks. The GTX 680M retains merit in scenarios where its higher pixel and texture rates matter, but the measured data gives the overall win to the newer Maxwell part.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 950M averages 8,135, while the NVIDIA GeForce GTX 680M averages 7,023, a difference of roughly 1,112 points in favor of the GTX 950M.

Q: How do the two GPUs compare in the Geekbench OpenCL test?

A: The GTX 950M scores 9,745 and the GTX 680M scores 9,230. The GTX 950M wins by 5.6% in this head-to-head measurement.

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

A: The GTX 680M has a 256-bit memory bus with GDDR5 memory, delivering 115.2 GB/s. The GTX 950M uses a 128-bit bus with DDR3 memory, providing 28.80 GB/s. The GTX 680M has a fourfold bandwidth advantage.

Q: Which GPU has more shading units and texture mapping units?

A: The GTX 680M has 1,344 shading units and 112 TMUs. The GTX 950M has 640 shading units and 40 TMUs. The GTX 680M has more than double the shading units and nearly triple the TMUs.

Q: Are both GPUs from the same architecture?

A: No. The GTX 950M uses the Maxwell architecture (chip GM107), while the GTX 680M uses the Kepler architecture (chip GK104). Both are built on a 28 nm process at TSMC.

Q: What is the transistor count difference?

A: The GTX 680M contains 3,540 million transistors on a 294 mm² die. The GTX 950M contains 1,870 million transistors on a 148 mm² die. The GTX 680M has nearly twice the transistor count and roughly double the die area.

Architecture Differences

The GTX 950M and GTX 680M come from different NVIDIA architecture generations. The GTX 950M is built on Maxwell, using the GM107 chip, while the GTX 680M uses Kepler with the GK104 chip. Both are manufactured by TSMC on a 28 nm process, but the similarities end there.

The GTX 680M packs 3,540 million transistors into a 294 mm² die, giving a transistor density of 12.0 million per square millimeter. The GTX 950M uses 1,870 million transistors on a 148 mm² die, achieving a slightly higher density of 12.6 million per square millimeter. Despite the older architecture, the GTX 680M’s larger chip delivers substantially more raw compute resources.

The GTX 680M features 1,344 shading units, 112 texture mapping units, and 32 raster output units. The GTX 950M has 640 shading units, 40 TMUs, and 16 ROPs. This means the GTX 680M has 110% more shading units, 180% more TMUs, and double the ROPs. The architectural efficiency of Maxwell, however, allows the GTX 950M to close much of the performance gap in real-world measurements.

Both GPUs support DirectX 12 (11_0) and OpenGL 4.6. The Vulkan support differs: the GTX 950M supports Vulkan 1.4, while the GTX 680M supports Vulkan 1.2.175. Neither GPU includes ray tracing cores or tensor cores.

The GTX 950M belongs to the GeForce 900M generation and follows the GeForce 800M line. The GTX 680M is part of the GeForce 600M generation, succeeding the GeForce 500M and preceding the GeForce 700M. The GTX 950M’s successor is the GeForce 10 Mobile series. The release dates reflect this generational gap: the GTX 680M launched in June 2012, while the GTX 950M arrived in March 2015.

Specification Differences

The two GPUs differ across nearly every specification field. The GTX 950M operates at a base clock of 993 MHz with a boost of 1124 MHz. The GTX 680M runs at a base of 719 MHz with a boost of 758 MHz, meaning the GTX 950M has a 274 MHz base clock advantage and a 366 MHz boost advantage.

Memory configuration shows a significant divergence. The GTX 950M has 4 GB of DDR3 on a 128-bit bus, with memory running at 900 MHz (1800 Mbps effective) and bandwidth of 28.80 GB/s. The GTX 680M also has 4 GB but uses GDDR5 on a 256-bit bus, with memory at 900 MHz (3.6 Gbps effective) and bandwidth of 115.2 GB/s. The GTX 680M’s bandwidth is four times higher.

Compute rates favor different cards. The GTX 950M achieves a pixel rate of 17.98 GPixel/s and texture rate of 44.96 GTexel/s. The GTX 680M achieves 21.22 GPixel/s and 84.90 GTexel/s. In floating-point performance, the GTX 680M reaches 2.038 TFLOPS, while the GTX 950M delivers 1,438.7 GFLOPS.

Power consumption differs meaningfully. The GTX 950M has a TDP of 75 W and uses an IGP slot width with no power connectors. The GTX 680M has a TDP of 100 W, uses an MXM Module slot width, and also has no power connectors. The bus interface differs as well: the GTX 950M uses PCIe 3.0 x8, while the GTX 680M uses MXM-B (3.0). Both list display outputs as portable device dependent.

The GTX 950M has a Geekbench Vulkan score of 6,525, a test the GTX 680M does not have recorded. The GTX 680M has a Geekbench Metal score of 4,815, a test the GTX 950M does not have recorded. Neither card has a launch MSRP in the database.

Head-to-Head Benchmarks

The database contains one direct head-to-head benchmark between these two GPUs: Geekbench OpenCL. The GTX 950M scores 9,745, and the GTX 680M scores 9,230. The GTX 950M wins by 5.6%, a margin that reflects the newer architecture’s efficiency despite the GTX 680M’s larger silicon.

This 515-point gap in raw score is meaningful when viewed against each card’s nearest rivals. The GTX 950M’s average of 8,135 places it within 0.4% of the NVIDIA GeForce GTX 980 (8,167) and 0.1% of the AMD Radeon R9 M360 (8,129). The GTX 680M’s average of 7,023 puts it 0.2% behind the NVIDIA T600 (7,035) and 1.1% ahead of the NVIDIA GeForce GTX 675M (6,946). The GTX 950M operates in a performance bracket roughly 1,100 points higher than the GTX 680M on average.

In the OpenCL test specifically, the GTX 950M’s 9,745 exceeds its own average by about 20%, while the GTX 680M’s 9,230 exceeds its average by about 31%. This indicates that OpenCL workloads are a relative strength for both cards, but the GTX 950M still holds the edge.

The GTX 680M’s theoretical advantages do not translate into a benchmark win. Its 115.2 GB/s bandwidth, 84.90 GTexel/s texture rate, and 2.038 TFLOPS are all substantially higher than the GTX 950M’s figures. Yet the recorded OpenCL result shows the GTX 950M ahead. The Maxwell architecture’s higher clock speeds and better compute efficiency per shading unit appear to compensate for the smaller resource pool.

Where Each One Wins

The GTX 950M wins the only direct benchmark comparison, so it is the clear choice for OpenCL compute tasks and general GPU-accelerated workloads. Its Geekbench OpenCL score of 9,745 and average benchmark score of 8,135 both exceed the GTX 680M’s corresponding figures. The GTX 950M also supports Vulkan 1.4, which may offer broader compatibility with modern Vulkan applications compared to the GTX 680M’s Vulkan 1.2.175 support.

The GTX 680M holds advantages in several specification areas that could matter for specific workloads, even though no head-to-head test confirms a win. Its 115.2 GB/s memory bandwidth is 300% higher than the GTX 950M’s 28.80 GB/s, which could benefit texture-heavy or memory-intensive operations. Its pixel rate of 21.22 GPixel/s and texture rate of 84.90 GTexel/s exceed the GTX 950M’s 17.98 GPixel/s and 44.96 GTexel/s, suggesting stronger fill-rate performance in rasterization-heavy scenarios.

The GTX 680M also has a higher TDP at 100 W versus 75 W, which may indicate a higher sustained power envelope for demanding loads. Its MXM Module form factor differs from the GTX 950M’s IGP design, meaning the two cards target different laptop chassis configurations.

For users whose primary metric is the recorded database scores, the GTX 950M is the safer recommendation. For workloads that depend heavily on memory bandwidth or raw texture throughput, the GTX 680M’s specification sheet makes a case, but the absence of a benchmark win in the database means that advantage remains theoretical rather than measured.