NVIDIA GeForce GTX 680M

NVIDIA’s GeForce GTX 680M represents the Kepler architecture’s push into high-end mobile graphics, built on a 28 nm process at TSMC with 3,540 million transistors on a 294 mm² die. The data shows a mobile part aimed at flagship laptops of its generation, and its benchmark scores place it in a tight cluster with several desktop and mobile contemporaries. With an average benchmark score of 7,193, the GTX 680M sits at the 38th percentile among all GPUs, indicating a solidly mid-to-upper-tier performer rather than a class leader.

Benchmark Performance

The GTX 680M delivers an average benchmark score of 7,193, derived from its Geekbench results: 9,571 in OpenCL and 4,815 in Metal. These figures immediately contextualize its position—it is not a top-tier part by modern standards, but it holds its own against a specific set of rivals. The most direct comparison is with the AMD Radeon Vega 8 Mobile, which scores 7,203, a delta of just -0.1%. This means the GTX 680M trails the Vega 8 by a negligible margin, effectively a statistical tie. Similarly, the NVIDIA GeForce GTX 750 scores 7,203, again a -0.1% delta, reinforcing that the 680M performs on par with that desktop card despite being a mobile part from an earlier era.

The picture shifts slightly when compared to the Quadro 5000, which posts an average score of 7,315. Here the GTX 680M is behind by 1.7%, a modest gap that suggests the Quadro’s workstation-oriented optimization gives it a small edge in compute-heavy tasks. Conversely, the NVIDIA T600 scores 7,068, and the GTX 680M leads it by 1.8%. This positive delta indicates that the older mobile chip still outperforms a newer entry-level workstation card in raw benchmark terms. The data implies a performance plateau: the GTX 680M sits within a ±2% band of four rivals, making its real-world differentiation dependent on driver support and specific workloads rather than raw compute dominance.

In terms of raw throughput, the GTX 680M’s 2.038 TFLOPS of FP32 performance, combined with a pixel rate of 21.22 GPixel/s and a texture rate of 84.90 GTexel/s, explains why it remains competitive. The 1,344 shading units and 112 texture mapping units are substantial for a 100 W mobile design, but the 32 ROPs are a limiting factor for high-resolution fill-rate demands, which becomes evident in the ray tracing and feature set analysis below.

Ray Tracing and Feature Set

The GTX 680M has no dedicated ray tracing cores and no tensor cores, as these are absent from the FACT PACK specifications. This is a Kepler-generation GPU, and its feature set reflects that era’s design priorities. The absence of RT and tensor hardware means that any ray-traced workloads, if attempted, would rely on compute shaders, which is inefficient compared to dedicated hardware on later architectures. The data does not list any ray tracing performance scores, so no quantitative claims can be made about its RT capabilities—only that the hardware lacks the specialized units.

API support, however, is broader than one might expect from a 2012-era part. The GTX 680M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Vulkan 1.2.175 support is notable because it enables modern cross-platform titles to run on this GPU, even if the underlying hardware is dated. DirectX 12 (11_0) is a feature-level limitation, meaning it does not support the full DirectX 12 feature set—only the 11_0 subset—which caps its ability to use certain modern rendering techniques like variable rate shading or mesh shaders. OpenGL 4.6 is fully supported, which is relevant for professional applications and emulators.

The lack of tensor cores also means no DLSS or AI-assisted upscaling, and the lack of RT cores means no hardware-accelerated ray tracing. The feature set is therefore a mix of modern API compatibility and outdated fixed-function hardware. This makes the GTX 680M viable for older titles and low-to-mid settings in contemporary games, but it cannot leverage the latest graphical advancements. The data suggests a GPU that is API-forward but hardware-limited, a combination that yields surprising longevity in esports titles but not in AAA releases.

How It Compares

vs. AMD Radeon Vega 8 Mobile: The GTX 680M scores 7,193 against Vega 8 Mobile’s 7,203, a delta of -0.1%. This is essentially a dead heat, with less than one-tenth of a percent separating them. The Vega 8 is a newer integrated solution, yet the dedicated GTX 680M manages to match it in average benchmark score. This suggests that the 680M’s dedicated memory and higher power envelope offset the Vega 8’s architectural efficiency gains. In practice, the GTX 680M might fare better in sustained loads due to its 100 W TDP, while the Vega 8 would throttle in thin-and-light chassis.

vs. NVIDIA GeForce GTX 750: Again, a -0.1% delta, with the GTX 750 scoring 7,203. The GTX 750 is a desktop card, but the 680M matches it exactly in average score. This is a strong indicator of the 680M’s efficiency, as it achieves desktop-class performance at a mobile power budget. The GTX 750 likely has higher clock stability due to better cooling, but the 680M’s 4 GB VRAM gives it an advantage in texture-heavy scenes. Benchmark results show parity, so the choice between them would come down to platform (laptop vs. desktop) rather than performance.

vs. NVIDIA Quadro 5000: The Quadro 5000 leads with 7,315, a 1.7% advantage over the GTX 680M. This is a modest but consistent gap, likely due to the Quadro’s professional driver optimizations for compute workloads. The GTX 680M’s 2.038 TFLOPS is close to the Quadro’s performance, but the Quadro’s higher average score suggests better sustained throughput in Geometry or OpenCL tasks. For gaming, the GTX 680M would likely outperform the Quadro due to game-specific drivers, but the data only shows average benchmarks.

vs. NVIDIA T600: The GTX 680M leads the T600 by 1.8%, with scores of 7,193 and 7,068, respectively. The T600 is a newer, low-profile workstation card, yet it falls behind the older mobile chip. This is a surprising result that implies the GTX 680M’s wider memory bus (256 bit vs. presumably narrower on the T600) and higher shading unit count (1,344) give it an edge in raw compute. The T600’s advantage would be in power efficiency and driver support, but the benchmark data clearly favors the GTX 680M.

Who Should Consider It

Benchmark results indicate the GTX 680M is best suited for 1080p gaming at medium to high settings in titles from its era (early-to-mid 2010s). Its 2.038 TFLOPS of FP32 performance and 115.2 GB/s memory bandwidth are sufficient for older games, but the 32 ROPs and lack of modern features like DLSS or ray tracing mean it struggles with 1440p or 4K resolutions. At 1080p, the GTX 680M can maintain playable framerates in esports titles (e.g., CS:GO, League of Legends) due to its high texture rate, but modern AAA games at high settings will likely push it beyond its limits.

For 1440p, the data suggests the GTX 680M is not a viable option, as its pixel rate of 21.22 GPixel/s and 115.2 GB/s bandwidth become bottlenecks. Users with 1080p displays and a tolerance for medium settings in newer games, or high settings in older ones, would find this GPU acceptable. It is also a candidate for legacy system builds, given its MXM-B (3.0) form factor and end-of-life status. However, the 38th percentile ranking indicates that most modern GPUs outperform it, so it should only be considered for retro gaming or light productivity workloads, not as a primary gaming solution.

FAQ

Q: How does the GTX 680M perform in modern games?

A: The GTX 680M has an average benchmark score of 7,193, placing it at the 38th percentile. This means it outperforms roughly 38% of all GPUs, which is insufficient for modern AAA titles at high settings, but it can handle esports and older titles at 1080p medium-to-high settings.

Q: Does the GTX 680M support ray tracing?

A: No. The FACT PACK lists no ray tracing cores and no tensor cores, so hardware-accelerated ray tracing is not available. Its DirectX 12 (11_0) support also limits the use of DXR-based features.

Q: What is the GTX 680M’s closest rival?

A: The AMD Radeon Vega 8 Mobile and NVIDIA GeForce GTX 750 both score 7,203, which is -0.1% relative to the GTX 680M’s 7,193. This makes them statistically identical in average benchmark performance.

Q: Can the GTX 680M handle 4K gaming?

A: The data indicates no. With 32 ROPs, a pixel rate of 21.22 GPixel/s, and 115.2 GB/s memory bandwidth, 4K resolutions would exceed its fill-rate and bandwidth capabilities, resulting in low framerates even in older titles.

Q: What APIs does the GTX 680M support?

A: It supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. This allows modern API-based games to run, but the DirectX 12 feature level is limited to 11_0.

Q: Is the GTX 680M faster than the NVIDIA T600?

A: Yes, the GTX 680M scores 7,193, which is 1.8% higher than the T600’s 7,068. This leads the older mobile chip over the newer workstation card in average benchmark performance.

Memory Subsystem

The GTX 680M is equipped with 4 GB of GDDR5 memory on a 256-bit bus, yielding a bandwidth of 115.2 GB/s. The memory clock is listed as 900 MHz with an effective data rate of 3.6 Gbps. This configuration is substantial for a mobile GPU of its generation, providing ample capacity for high-resolution textures. The 256-bit bus is wider than many contemporary mobile parts, which helps mitigate the bandwidth bottleneck in memory-intensive scenes.

For high resolutions, the 4 GB capacity is sufficient for 1080p and even 1440p textures in older games, but the 115.2 GB/s bandwidth is the limiting factor. Compare this to modern GPUs with 300+ GB/s bandwidth, and the GTX 680M will struggle with 4K textures or heavy anti-aliasing. The 32 ROPs further constrain high-resolution pixel throughput, so while the memory capacity is generous, the bandwidth and ROP count prevent the GTX 680M from being a high-resolution performer. The data suggests that 1080p is the sweet spot, where the 115.2 GB/s bandwidth can feed the 2.038 TFLOPS compute without stalling.

Power and Cooling

The GTX 680M has a thermal design power (TDP) of 100 W, which is modest for a GPU with 1,344 shading units and a 256-bit memory bus. This power envelope allows for thinner laptop designs, but it also means that sustained boost clocks (758 MHz) may be difficult to maintain under heavy load without adequate cooling. The base clock is 719 MHz, and the boost clock adds only 39 MHz, indicating a conservative overclocking headroom.

The slot width is listed as MXM Module, and the bus interface is MXM-B (3.0), meaning it is a replaceable module rather than a soldered chip. The power connectors are listed as "None," which suggests that power is supplied through the MXM slot itself, not external PCIe power cables. There is no suggested PSU listed in the FACT PACK, so no recommendation can be made regarding power supply units. The 100 W TDP means that a laptop cooling solution with a decent heat pipe and fan should suffice, but users should be aware that the GTX 680M is an end-of-life product, and replacement cooling parts may be hard to find. The absence of power connectors simplifies installation but also means there is no headroom for overclocking via external power.