NVIDIA GeForce GTX 680 vs NVIDIA GeForce RTX 3060 Ti Comparison

Head-to-Head Benchmarks

The recorded data shows a decisive victory for the NVIDIA GeForce RTX 3060 Ti across the two shared benchmark tests. In Geekbench OpenCL, the RTX 3060 Ti scores 78,927 points against the GTX 680’s 16,906 points, a 366.9% delta. This is not a marginal generational improvement; it is a fundamental leap in raw compute throughput. The OpenCL test exercises general-purpose GPU compute, and the RTX 3060 Ti’s score places it in a completely different performance tier.

In Geekbench Vulkan, the gap narrows relative to OpenCL but remains overwhelming. The RTX 3060 Ti records 47,784 points, while the GTX 680 manages 17,646 points, a 170.8% delta. Vulkan is a lower-level graphics API that can reduce CPU overhead and expose more of the GPU’s raw hardware capabilities. Even with that efficiency advantage, the GTX 680 cannot approach the RTX 3060 Ti’s output. The delta between the two tests is telling: OpenCL shows a 366.9% difference, while Vulkan shows a 170.8% difference, indicating that the GTX 680’s architecture scales relatively better in graphics-oriented workloads than in pure compute, but still falls far short.

The win count is clear: the RTX 3060 Ti wins 2 out of 2 head-to-head benchmarks, while the GTX 680 wins none. The database’s average benchmark score metric reinforces this. The RTX 3060 Ti has an average score of 16,129 across all recorded benchmarks, while the GTX 680 averages 14,150. That is a 14% difference in the aggregate, though the head-to-head deltas are much larger because they isolate specific tests rather than mixing in other workload types. The RTX 3060 Ti also holds a 59th percentile ranking among all GPUs in the database, compared to the GTX 680’s 55th percentile. Both are mid-pack performers in the broader GPU landscape, but the RTX 3060 Ti sits noticeably higher.

FAQ

Q: Which card wins in Geekbench OpenCL, and by how much?

A: The NVIDIA GeForce RTX 3060 Ti wins with a score of 78,927 versus the GTX 680’s 16,906, a 366.9% advantage.

Q: Is the Vulkan result closer than the OpenCL result?

A: Yes. In Geekbench Vulkan, the RTX 3060 Ti scores 47,784 against the GTX 680’s 17,646, a 170.8% delta, which is less than half the percentage gap seen in OpenCL.

Q: What is the average benchmark score for each card?

A: The RTX 3060 Ti has an average score of 16,129 across all benchmarks in the database, while the GTX 680 has an average of 14,150.

Q: How do these cards rank against all GPUs in the database?

A: The RTX 3060 Ti sits at the 59th percentile, and the GTX 680 sits at the 55th percentile.

Q: Which card has the higher transistor count?

A: The RTX 3060 Ti has 17,400 million transistors on a 392 mm² die, while the GTX 680 has 3,540 million transistors on a 294 mm² die. The transistor density is 44.4 million per mm² for the RTX 3060 Ti versus 12.0 million per mm² for the GTX 680.

Q: Are there any benchmarks where the GTX 680 wins?

A: No. The head-to-head results show the RTX 3060 Ti wins both recorded tests, and the GTX 680 has zero wins in the comparison.

Where Each One Wins

The RTX 3060 Ti wins in every recorded workload category. In compute-heavy tasks, as shown by Geekbench OpenCL, its 366.9% lead means it is the only reasonable choice for GPU-accelerated workloads like rendering, scientific simulation, or machine learning preprocessing. The FP32 throughput of 16.20 TFLOPS versus 3.250 TFLOPS reinforces this: the RTX 3060 Ti processes four times the floating-point operations per second. In graphics-oriented workloads, the Vulkan result shows a 170.8% lead, which translates to smoother frame rates and higher detail settings in modern games that use Vulkan or DirectX 12.

The GTX 680 has no scenario where the recorded data favors it. Its only advantages are qualitative: it has a lower transistor density (12.0M / mm² versus 44.4M / mm²), which historically may imply simpler manufacturing, and it uses a 28 nm process from TSMC versus the 8 nm Samsung process in the RTX 3060 Ti. Neither of these translates into a performance win. The GTX 680’s power draw of 195 W is only 5 W lower than the RTX 3060 Ti’s 200 W, so it does not even offer a meaningful efficiency advantage. The GTX 680 is an older design from 2012, and the data confirms it is outclassed across the board.

For users with legacy software that only supports older graphics APIs, the GTX 680 might still function, but the RTX 3060 Ti supports DirectX 12 Ultimate (12_2) versus the GTX 680’s DirectX 12 (11_0), so the newer card is forward-compatible with modern titles. In short, the RTX 3060 Ti is the winner in every measured dimension.

Specification Differences

The two cards differ in nearly every specification. The RTX 3060 Ti uses 8 GB of GDDR6 memory on a 256-bit bus, with bandwidth of 448.0 GB/s. The GTX 680 uses 2 GB of GDDR5 memory on a 256-bit bus, with bandwidth of 192.3 GB/s. The memory capacity difference is 4x, and the bandwidth difference is 2.3x. The RTX 3060 Ti has 4,864 shading units, 152 texture mapping units, and 80 render output units. The GTX 680 has 1,536 shading units, 128 TMUs, and 32 ROPs. That is a 3.2x difference in shading units and a 2.5x difference in ROPs.

Clock speeds differ significantly. The RTX 3060 Ti runs at a base clock of 1410 MHz and a boost clock of 1665 MHz. The GTX 680 runs at 1006 MHz base and 1058 MHz boost. Memory clocks are 1750 MHz (14 Gbps effective) for the RTX 3060 Ti versus 1502 MHz (6 Gbps effective) for the GTX 680. The pixel rate is 133.2 GPixel/s for the RTX 3060 Ti versus 33.86 GPixel/s for the GTX 680. The texture rate is 253.1 GTexel/s versus 135.4 GTexel/s. FP32 compute is 16.20 TFLOPS versus 3.250 TFLOPS. The RTX 3060 Ti also supports FP16 at 16.20 TFLOPS (1:1), while the GTX 680 has no recorded FP16 capability.

Power and connectivity also differ. The RTX 3060 Ti has a TDP of 200 W and requires a 550 W suggested PSU, with a single 12-pin power connector. The GTX 680 has a TDP of 195 W, a 450 W suggested PSU, and two 6-pin connectors. The bus interface is PCIe 4.0 x16 for the RTX 3060 Ti versus PCIe 3.0 x16 for the GTX 680. Display outputs are 1x HDMI 2.1 and 3x DisplayPort 1.4a on the RTX 3060 Ti, versus 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2 on the GTX 680. Physical dimensions: the RTX 3060 Ti is 242 mm long and 112 mm high, while the GTX 680 is 256 mm long, 111 mm high, and 38 mm wide. The RTX 3060 Ti has a dual-slot width, as does the GTX 680.

Architecture Differences

The RTX 3060 Ti is built on the Ampere architecture with the GA104 chip, fabricated on Samsung’s 8 nm process. The GTX 680 uses the Kepler architecture with the GK104 chip, fabricated on TSMC’s 28 nm process. The transistor count is 17,400 million versus 3,540 million, a 4.9x difference. Die size is 392 mm² versus 294 mm², so the RTX 3060 Ti packs far more transistors into a larger but denser die. Transistor density is 44.4 million per mm² versus 12.0 million per mm², a 3.7x difference.

The RTX 3060 Ti includes 38 ray tracing cores and 152 tensor cores, neither of which exists on the GTX 680. These hardware units enable real-time ray tracing and AI-accelerated features like DLSS, which the GTX 680 cannot perform. The RTX 3060 Ti supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The GTX 680 supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The API support difference matters for modern game compatibility, as DirectX 12 Ultimate includes features like mesh shaders and variable rate shading that require the newer hardware.

The memory technology also differs architecturally. GDDR6 on the RTX 3060 Ti operates at 14 Gbps effective, while GDDR5 on the GTX 680 operates at 6 Gbps effective. The RTX 3060 Ti’s FP16 support at 1:1 ratio with FP32 is a significant architectural feature, as it allows compute-heavy workloads to double throughput. The GTX 680 has no FP16 path in the recorded data. The RTX 3060 Ti’s release date is 2020-11-30, while the GTX 680’s is 2012-03-21, an eight-year gap that explains the architectural generational leap.

The Verdict

The data is unambiguous. The NVIDIA GeForce RTX 3060 Ti outperforms the NVIDIA GeForce GTX 680 in every recorded benchmark, with deltas ranging from 170.8% in Vulkan to 366.9% in OpenCL. The RTX 3060 Ti has a higher average benchmark score (16,129 versus 14,150), a higher percentile ranking (59th versus 55th), and more than four times the shading units, FP32 throughput, and memory bandwidth. It also adds ray tracing cores, tensor cores, and modern API support that the GTX 680 lacks entirely.

For anyone building a system for modern gaming or compute workloads, the RTX 3060 Ti is the only logical choice. The GTX 680’s 2 GB of memory is insufficient for current game textures, and its lack of hardware ray tracing and tensor cores means it cannot run contemporary features. The RTX 3060 Ti’s 8 GB of GDDR6 memory and 448.0 GB/s bandwidth provide a comfortable buffer for high-resolution textures and compute buffers.

The GTX 680 might appeal to collectors or retro-build enthusiasts who specifically want a Kepler-era card for legacy software. Its 195 W TDP is marginally lower than the RTX 3060 Ti’s 200 W, but that 5 W difference is negligible in practice. The GTX 680’s PCIe 3.0 interface and older display outputs (2x DVI, HDMI 1.4a, DisplayPort 1.2) limit its usefulness with modern monitors and motherboards. The RTX 3060 Ti’s PCIe 4.0 interface and DisplayPort 1.4a outputs are ready for current hardware.

The verdict: choose the RTX 3060 Ti for any workload that benefits from raw compute, modern graphics APIs, or ray tracing. Choose the GTX 680 only if the goal is historical preservation or running software that specifically requires Kepler-era hardware. The benchmark results show no scenario where the GTX 680 competes on performance. The RTX 3060 Ti is the definitive winner in this comparison.