NVIDIA GeForce GTX 680 vs NVIDIA T400 Comparison

Head-to-Head Benchmarks

The recorded data shows a close contest between these two cards, with each taking one benchmark win in the head-to-head comparison. In Geekbench OpenCL, the NVIDIA T400 scores 17,039 against the GTX 680's 16,906, a margin of 0.8% in favor of the T400. This is a narrow victory, and the delta sits within the noise of typical run-to-run variation. However, the Vulkan results tell a different story. The GTX 680 posts 17,646, while the T400 manages 15,976, giving the older card a decisive 9.5% advantage in that test.

Looking at the broader database context, the T400's average benchmark score is 16,508, which places it at the 60th percentile among all GPUs. Its nearest rivals include the NVIDIA GeForce RTX 5090 D V2 (16,504, a 0% delta), the AMD Radeon PRO W7500 (16,415, 0.6% behind), and the NVIDIA RTX PRO 6000 Blackwell (16,408, 0.6% behind). The T400 effectively sits in a cluster of cards with nearly identical average scores, meaning its positioning is tightly contested. The GTX 680, meanwhile, has an average benchmark score of 14,150, which puts it at the 55th percentile. Its closest comparators are the NVIDIA Tesla K10 (14,029, 0.9% lower), the NVIDIA GeForce GTX 1070 Ti (14,277, 0.9% higher), and the Intel Iris Xe MAX Graphics (14,315, 1.2% higher). The GTX 680's average is dragged down by its Metal score of 7,897, which is not a test the T400 runs, so the average comparison is not perfectly apples-to-apples.

When interpreting the head-to-head results, the Vulkan gap is the more substantial differentiator. A 9.5% lead for the GTX 680 suggests that in modern Vulkan workloads, the older architecture retains a meaningful edge. The OpenCL result, by contrast, is effectively a tie, with the T400's 0.8% lead being too small to draw strong conclusions from. The data indicates that neither card is a clear overall winner; instead, the choice depends on which API or workload matters more to the user.

The Verdict

Strictly from the recorded data, the NVIDIA GeForce GTX 680 is the better choice for users prioritizing Vulkan performance. Its 17,646 Vulkan score is 9.5% higher than the T400's 15,976, and that is a substantial margin in a modern graphics API. The GTX 680 also has a higher average benchmark score in the database at 14,150 versus 16,508 for the T400, but this comparison is complicated by the fact that the T400 does not have a Metal score, while the GTX 680 does. The GTX 680's Metal result is 7,897, which pulls its average down considerably. If the Metal score is excluded from consideration, the two cards are much closer in average performance.

The NVIDIA T400 is the better choice for users who need a single-slot, low-power solution. Its 30 W TDP, single-slot width, and lack of power connectors make it far easier to integrate into compact or densely populated systems. The T400 also supports newer display outputs with 3x mini-DisplayPort 1.4a, whereas the GTX 680 offers a mix of DVI, HDMI 1.4a, and DisplayPort 1.2. For users whose primary concern is Vulkan performance, the GTX 680 is the data-backed pick. For users who need modern display connectivity and minimal power draw, the T400 is the only sensible option between the two.

Architecture Differences

The two cards come from different NVIDIA architecture generations. The T400 is built on the Turing architecture, using the TU117 chip, while the GTX 680 uses the Kepler architecture with the GK104 chip. The T400 is fabricated on a 12 nm process at TSMC, whereas the GTX 680 uses a 28 nm process, also at TSMC. This process difference is significant: the T400 packs 4,700 million transistors into a 200 mm² die, giving a transistor density of 23.5 million per mm². The GTX 680 has 3,540 million transistors on a 294 mm² die, for a density of 12.0 million per mm². The T400's newer process allows more than double the transistor density per unit area.

The T400 belongs to the Quadro Turing generation, while the GTX 680 is part of the GeForce 600 generation. Neither card has ray tracing cores or tensor cores, as both predate or lack those dedicated units. The T400 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The GTX 680 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The Vulkan version difference is notable, as the T400 supports a newer specification. The T400 also has a higher transistor count despite the smaller die, which reflects the density advantage of the 12 nm node.

Specification Differences

The two cards differ across nearly every core specification. The T400 has 384 shading units, 24 texture mapping units, and 16 raster output pipelines. The GTX 680 has 1,536 shading units, 128 TMUs, and 32 ROPs. This is a 4x difference in shading units and a 5.3x difference in TMUs, giving the GTX 680 a massive theoretical compute advantage on paper. However, clock speeds tell a different story. The T400 has a base clock of 420 MHz and a boost clock of 1,425 MHz. The GTX 680 has a base clock of 1,006 MHz and a boost of 1,058 MHz. The T400's boost clock is 34.7% higher than the GTX 680's, which partially compensates for its lower core count.

Memory configurations are also very different. The T400 uses 2 GB of GDDR6 on a 64-bit bus, delivering 80.00 GB/s of bandwidth. The GTX 680 uses 2 GB of GDDR5 on a 256-bit bus, delivering 192.3 GB/s of bandwidth. The GTX 680's bandwidth is 2.4x higher, which is a significant advantage in bandwidth-sensitive workloads. The memory clock is 1,250 MHz (10 Gbps effective) for the T400 versus 1,502 MHz (6 Gbps effective) for the GTX 680. The effective data rate is higher on the T400, but the narrower bus limits total bandwidth.

Pixel and texture rates reflect the same pattern. The T400 achieves 22.80 GPixel/s and 34.20 GTexel/s. The GTX 680 achieves 33.86 GPixel/s and 135.4 GTexel/s. The GTX 680 is 48.5% faster in pixel fill rate and nearly 4x faster in texture fill rate. FP32 compute is 1,094.4 GFLOPS for the T400 and 3.250 TFLOPS for the GTX 680, making the GTX 680 about 3x faster in single-precision floating-point. The T400 does have FP16 capability at 2.189 TFLOPS (2:1 ratio), while the GTX 680 has no recorded FP16 performance.

Power and physical specifications diverge sharply. The T400 has a TDP of 30 W, is single-slot, and requires no power connectors, with a suggested PSU of 200 W. The GTX 680 has a TDP of 195 W, is dual-slot, requires 2x 6-pin power connectors, and needs a 450 W PSU. The GTX 680 is also physically larger at 256 mm in length, 111 mm in height, and 38 mm in width, while the T400 has no recorded dimensions. The T400's release date is May 5, 2021, while the GTX 680 launched on March 21, 2012. Both are end-of-life products. The GTX 680 has a recorded launch MSRP of 499 USD.

FAQ

Q: Which card has a higher average benchmark score?

A: The NVIDIA T400 has an average benchmark score of 16,508, which is higher than the GTX 680's 14,150. However, the GTX 680 has a Metal benchmark score of 7,897 that the T400 does not have, which lowers its average.

Q: Which card performs better in Vulkan?

A: The GTX 680 scores 17,646 in Geekbench Vulkan, while the T400 scores 15,976. The GTX 680 leads by 9.5% in this test.

Q: Which card has a lower power requirement?

A: The T400 has a TDP of 30 W and a suggested PSU of 200 W, while the GTX 680 has a TDP of 195 W and a suggested PSU of 450 W.

Q: What memory types do the two cards use?

A: The T400 uses 2 GB of GDDR6 with a 64-bit bus and 80.00 GB/s bandwidth. The GTX 680 uses 2 GB of GDDR5 with a 256-bit bus and 192.3 GB/s bandwidth.

Q: Which card has more shading units?

A: The GTX 680 has 1,536 shading units, while the T400 has 384 shading units. The GTX 680 also has 128 TMUs and 32 ROPs, versus 24 TMUs and 16 ROPs on the T400.

Q: What is the transistor density difference?

A: The T400 has a transistor density of 23.5 million per mm² on a 12 nm process, while the GTX 680 has 12.0 million per mm² on a 28 nm process.

Where Each One Wins

The NVIDIA T400 wins in scenarios that favor its architectural advantages. Its 12 nm Turing process delivers much higher transistor density, which translates to better efficiency per watt. The 30 W TDP and single-slot design make it the only practical choice for compact workstations, embedded systems, or any chassis where space and thermals are at a premium. The T400 also supports newer display outputs (3x mini-DisplayPort 1.4a) and a newer Vulkan specification (1.4 versus 1.2.175). In OpenCL, the T400 edges out the GTX 680 by 0.8%, so for compute workloads using that API, the two are effectively tied, but the T400's lower power draw means it achieves parity with far less energy.

The NVIDIA GeForce GTX 680 wins in raw performance metrics that matter for traditional rendering. Its 9.5% Vulkan lead is the clearest single-test advantage in the comparison. The GTX 680 also has 4x the shading units, 5.3x the TMUs, 2x the ROPs, and 2.4x the memory bandwidth of the T400. Its FP32 compute of 3.250 TFLOPS dwarfs the T400's 1,094.4 GFLOPS. Pixel rate is 48.5% higher and texture rate is nearly 4x higher. For users running Vulkan-based games or applications that leverage high texture throughput, the GTX 680 is the stronger card. Its 256-bit memory bus and 192.3 GB/s bandwidth give it a substantial edge in bandwidth-bound workloads.

In summary, the T400 is the efficiency pick for modern connectivity and low power draw, while the GTX 680 is the performance pick for Vulkan and raw throughput. The data does not declare an overall winner, as each card leads in different dimensions. Users should weigh the 9.5% Vulkan deficit against the 165 W TDP difference and the display output advantages of the T400.