NVIDIA GeForce GTX 675MX vs NVIDIA GeForce GTX 960M Comparison

Head-to-Head Benchmarks

The recorded data shows a single direct benchmark comparison between the NVIDIA GeForce GTX 960M and the NVIDIA GeForce GTX 675MX. In the Geekbench OpenCL test, the GTX 960M scores 11045, while the GTX 675MX scores 10723. This gives the GTX 960M a 3% advantage. The margin is modest, but it is a consistent win for the newer Maxwell-based part.

The overall average benchmark scores reinforce this picture. The GTX 960M posts an average score of 9645, placing it in the 46th percentile of all GPUs. The GTX 675MX averages 8427, which lands in the 43rd percentile. The difference in average scores is approximately 14.5% in favor of the GTX 960M, despite the fact that the GTX 675MX carries substantially more raw hardware resources. This is a case where architectural efficiency overcomes a deficit in sheer core count and memory bus width.

Looking at the nearest rivals for each card provides additional context. The GTX 960M sits in a tight cluster: the NVIDIA Quadro K5000 averages 9637 (0.1% behind), the AMD Radeon Pro WX 2100 averages 9653 (0.1% ahead), the NVIDIA Quadro P4000 averages 9665 (0.2% ahead), and the NVIDIA Tesla C2070 averages 9716 (0.7% ahead). The GTX 960M is effectively at parity with that group, trading tenths of a percent. The GTX 675MX occupies a slightly lower tier. Its closest rival, the AMD Radeon R9 M375X, averages 8325, which is 1.2% behind the GTX 675MX. The AMD Radeon 880M averages 8436 (0.1% ahead), the NVIDIA GeForce MX330 averages 8458 (0.4% ahead), and the AMD Radeon HD 8870M averages 8462 (0.4% ahead). The GTX 675MX is competitive in its own tier, but that tier sits roughly 1200 points lower than the GTX 960M's tier.

The head-to-head result is not dramatic, but it is unambiguous. The GTX 960M wins the only shared benchmark, and it wins the average-score comparison by a wider margin. The data indicates that the GTX 960M is the faster GPU overall, even though the GTX 675MX has more shading units, more texture mapping units, more render output units, and a wider memory bus.

Architecture Differences

The two GPUs come from different NVIDIA architectures and different generations. The GTX 960M is built on the Maxwell architecture, using the GM107 chip. It belongs to the GeForce 900M generation and was released on 2015-03-12. The GTX 675MX is built on the Kepler architecture, using the GK104 chip. It belongs to the GeForce 600M generation and was released on 2012-09-30. Both are manufactured on a 28 nm process at TSMC, but that is where the fabrication similarities end.

The transistor budgets differ significantly. The GM107 chip in the GTX 960M contains 1,870 million transistors on a die size of 148 mm². The GK104 chip in the GTX 675MX contains 3,540 million transistors on a die size of 294 mm². The transistor density is nearly identical: 12.6 million transistors per mm² for the GTX 960M and 12.0 million per mm² for the GTX 675MX. The larger GK104 die packs nearly twice the transistors, which explains the GTX 675MX's higher core counts.

The shading unit counts tell a clear story. The GTX 960M has 640 shading units, 40 TMUs, and 16 ROPs. The GTX 675MX has 960 shading units, 80 TMUs, and 32 ROPs. In every compute and rasterization resource category, the GTX 675MX has more hardware. Yet the GTX 960M still wins the benchmark. The Maxwell architecture extracts more usable performance per transistor and per shading unit than Kepler does.

Memory configurations also diverge. The GTX 960M uses 4 GB of GDDR5 on a 128-bit bus, with memory clocked at 1253 MHz and 5 Gbps effective. This produces a bandwidth of 80.19 GB/s. The GTX 675MX uses 2 GB of GDDR5 on a 256-bit bus, with memory clocked at 900 MHz and 3.6 Gbps effective. This produces a bandwidth of 115.2 GB/s. The GTX 675MX has a 44% bandwidth advantage, but the GTX 960M has twice the memory capacity. In practice, the wider bus does not translate into a benchmark win for the older card.

Clock behavior also differs. The GTX 960M has a base clock of 1097 MHz and a boost clock of 1176 MHz. The GTX 675MX has no recorded base or boost clock in the database, leaving its clock behavior unspecified. The GTX 960M's boost capability is a key part of its efficiency story.

The compute rate measurements reflect the architectural gap. The GTX 960M delivers 1.505 TFLOPS of FP32 performance, a pixel rate of 18.82 GPixel/s, and a texture rate of 47.04 GTexel/s. The GTX 675MX delivers 1,255.7 GFLOPS (approximately 1.256 TFLOPS), a pixel rate of 13.08 GPixel/s, and a texture rate of 52.32 GTexel/s. The GTX 960M wins in FP32 and pixel rate, while the GTX 675MX wins in texture rate. The power draw favors the newer card: the GTX 960M has a TDP of 75 W, while the GTX 675MX is rated at 100 W.

API support is broadly similar, with one notable difference. Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 960M supports Vulkan 1.4, while the GTX 675MX supports Vulkan 1.2.175. The newer card has a more recent Vulkan specification.

Other characteristics are shared. Both use an MXM-B (3.0) bus interface, both are MXM modules, both have no power connectors, and both have display outputs described as portable device dependent. Both are end-of-life products.

FAQ

Q: Which GPU wins the only direct benchmark comparison?

A: The NVIDIA GeForce GTX 960M wins the Geekbench OpenCL test with a score of 11045 versus 10723 for the GTX 675MX, a 3% advantage.

Q: How do the average benchmark scores compare?

A: The GTX 960M averages 9645, placing it in the 46th percentile of all GPUs. The GTX 675MX averages 8427, placing it in the 43rd percentile. The GTX 960M is ahead by roughly 14.5%.

Q: Does the GTX 675MX have more raw hardware resources?

A: Yes. The GTX 675MX has 960 shading units, 80 TMUs, and 32 ROPs, compared to 640 shading units, 40 TMUs, and 16 ROPs for the GTX 960M. It also has a 256-bit memory bus versus a 128-bit bus.

Q: Why does the GTX 960M win despite having fewer resources?

A: The GTX 960M uses the Maxwell architecture, which delivers higher FP32 throughput (1.505 TFLOPS versus 1,255.7 GFLOPS) and a higher pixel rate (18.82 GPixel/s versus 13.08 GPixel/s) while drawing less power (75 W versus 100 W). The GTX 675MX only leads in texture rate (52.32 GTexel/s versus 47.04 GTexel/s) and memory bandwidth (115.2 GB/s versus 80.19 GB/s).

Q: What is the difference in memory capacity and bandwidth?

A: The GTX 960M has 4 GB of GDDR5 on a 128-bit bus with 80.19 GB/s of bandwidth. The GTX 675MX has 2 GB of GDDR5 on a 256-bit bus with 115.2 GB/s of bandwidth.

Q: Which card has better API support?

A: Both support DirectX 12 (11_0) and OpenGL 4.6. The GTX 960M supports Vulkan 1.4, while the GTX 675MX supports Vulkan 1.2.175, so the GTX 960M has the newer Vulkan specification.

Specification Differences

| Specification | NVIDIA GeForce GTX 960M | NVIDIA GeForce GTX 675MX |

|---|---|---|

| Chip | GM107 | GK104 |

| Architecture | Maxwell | Kepler |

| Generation | GeForce 900M | GeForce 600M |

| Release Date | 2015-03-12 | 2012-09-30 |

| Transistors | 1,870 million | 3,540 million |

| Die Size | 148 mm² | 294 mm² |

| Transistor Density | 12.6M / mm² | 12.0M / mm² |

| Base Clock | 1097 MHz | Not recorded |

| Boost Clock | 1176 MHz | Not recorded |

| Memory Clock | 1253 MHz, 5 Gbps effective | 900 MHz, 3.6 Gbps effective |

| Memory Size | 4 GB | 2 GB |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 80.19 GB/s | 115.2 GB/s |

| Shading Units | 640 | 960 |

| TMUs | 40 | 80 |

| ROPs | 16 | 32 |

| Pixel Rate | 18.82 GPixel/s | 13.08 GPixel/s |

| Texture Rate | 47.04 GTexel/s | 52.32 GTexel/s |

| FP32 | 1.505 TFLOPS | 1,255.7 GFLOPS |

| TDP | 75 W | 100 W |

| Vulkan | 1.4 | 1.2.175 |

| Predecessor | GeForce 800M | GeForce 500M |

| Successor | GeForce 10 Mobile | GeForce 700M |

The Verdict

The data points to the NVIDIA GeForce GTX 960M as the stronger GPU. It wins the only shared benchmark, the Geekbench OpenCL test, by 3%. It holds a substantially higher average benchmark score, 9645 versus 8427, a difference of roughly 14.5%. It also occupies a higher percentile rank among all GPUs, 46th versus 43rd.

The GTX 675MX cannot overcome its architectural disadvantages. It has more shading units, more TMUs, more ROPs, and a wider memory bus, but it still trails in compute performance, pixel throughput, and overall benchmark results. Its higher TDP of 100 W, compared to 75 W for the GTX 960M, means it consumes more power while delivering less performance.

For any workload that relies on general GPU compute or OpenCL performance, the GTX 960M is the clear pick. The only areas where the GTX 675MX shows a measurable lead are texture rate and memory bandwidth, but these advantages do not translate into a benchmark victory. The GTX 960M also offers double the memory capacity, 4 GB versus 2 GB, which is relevant for larger datasets and texture-heavy scenarios.

The GTX 675MX is an older design from the Kepler era, released roughly 2.5 years earlier. It belongs to a previous generation, the GeForce 600M series, and its successor is the GeForce 700M series. The GTX 960M belongs to the GeForce 900M series and was succeeded by GeForce 10 Mobile. The newer product simply performs better across the board.

Where Each One Wins

The NVIDIA GeForce GTX 960M wins in every benchmark category that matters for overall GPU performance. The Geekbench OpenCL score of 11045 is the headline result. The average benchmark score of 9645 is higher than the GTX 675MX's 8427. The FP32 throughput of 1.505 TFLOPS exceeds the GTX 675MX's 1,255.7 GFLOPS. The pixel rate of 18.82 GPixel/s beats 13.85 GPixel/s. The GTX 960M is the choice for users prioritizing compute performance, modern API support, and memory capacity. Its 4 GB frame buffer doubles the 2 GB of the GTX 675MX. Its 75 W TDP makes it the more power-efficient option. Shading-heavy workloads, general 3D acceleration and OpenCL-based applications should favor the GTX 960M.

The NVIDIA GeForce GTX 675MX claims only niche victories. Its memory bandwidth of 115.2 GB/s is 44% higher than the GTX 960M's 80.19 GB/s, which could benefit bandwidth-bound tasks that fit within its 2 GB capacity. Its texture rate of 52.32 GTexel/s is ahead of the GTX 960M's 47.04 GTexel/s. The larger 256-bit memory bus and higher TMU count are the only hardware advantages the GTX 675MX can point to. The GTX 675MX also has the higher transistor count, 3,540 million versus 1,870 million, and the larger die at 294 mm² versus 148 mm², but these are not performance advantages in the recorded benchmarks.

The realistic conclusion is that the GTX 675MX is outclassed in practice. The GTX 960M wins the head-to-head comparison, wins the average-score comparison, and does so with lower power consumption. Users who find themselves with a GTX 675MX should expect slightly below the performance tier of the GTX 960M, roughly 14.5% based on average scores. Users with a GTX 960M receive the better overall package, even if the GTX 675MX retains a few theoretical hardware strengths that do not surface in the benchmark data. For compute and general use, the GTX 960M is the recommended option. The GTX 675MX is only preferable where memory bandwidth is the absolute bottleneck and texture throughput is critical, conditions that are rare in the recorded tests.