NVIDIA GeForce GTX 670 vs NVIDIA GeForce GTX 770 Comparison

Head-to-Head Benchmarks

The recorded data shows a clean sweep for the NVIDIA GeForce GTX 770 across all three benchmark tests. The largest margin comes in Geekbench Metal, where the GTX 770 scores 9369 against the GTX 670's 7424, a 26.2% advantage. This is a substantial gap, indicating that the newer card holds a significant edge in compute workloads that leverage Apple's Metal API.

The Vulkan test tells a similar story, though with slightly less drama. The GTX 770 posts 19166 points versus 15553 for the GTX 670, a 23.2% delta. That is a decisive win in a modern cross-platform graphics API, and it suggests the GTX 770 will feel noticeably faster in current games and applications that use Vulkan.

The OpenCL result is far closer. Here the GTX 770 scores 15707 against 15341 for the GTX 670, a modest 2.4% lead. While still a win for the newer card, this margin is small enough that day-to-day differences in OpenCL workloads would be difficult to notice outside of synthetic testing. The data indicates that both cards are roughly equivalent in raw OpenCL compute throughput, which makes sense given their shared architecture and similar core configurations.

Looking at the broader database context, the GTX 770 sits at the 57th percentile of all GPUs, with an average benchmark score of 14747. Its nearest rivals include the AMD Radeon RX 5500 XT (14692, a 0.4% gap) and the NVIDIA Quadro M2000 (14532, 1.5% behind). The GTX 770 also edges out the GeForce GTX 660 Ti by 2.1% and the AMD Radeon Pro 560X by 2.2%. These comparisons place the GTX 770 in a competitive mid-range position even against much newer hardware.

The GTX 670, by contrast, holds the 52nd percentile with an average score of 12773. Its closest peers are the GeForce GTX 590 (12830, GTX 670 is 0.4% behind), the AMD Radeon Pro 455 (12831, 0.4% behind), the AMD Radeon RX 7600M XT (12710, GTX 670 is 0.5% ahead), and the AMD FirePro W5100 (12847, 0.6% behind). These are all older or lower-tier parts, and the GTX 670's position among them reflects its status as a solid but not exceptional performer relative to its era.

The head-to-head win count is 3-0 in favor of the GTX 770. There is no benchmark in the database where the GTX 670 comes out ahead. That is a clean result, and it makes the performance hierarchy between these two cards unambiguous.

Architecture Differences

Both cards are built on the same GK104 chip, use the Kepler architecture, and are fabricated on TSMC's 28 nm process. Both pack 3,540 million transistors on a 294 mm² die, with a transistor density of 12.0M per mm². The fundamental silicon is identical; what separates them is how NVIDIA configured and clocked that silicon.

The GTX 770 enables more of the chip. It has 1536 shading units, 128 texture mapping units, and 32 ROPs. The GTX 670 is a cut-down version with 1344 shading units, 112 TMUs, and the same 32 ROPs. That is a 14.3% difference in shading units and a 14.3% difference in TMUs, which directly explains part of the performance gap. Both cards use 2 GB of GDDR5 memory on a 256-bit bus, but the GTX 770 runs its memory at 1753 MHz (7 Gbps effective) for 224.4 GB/s of bandwidth, while the GTX 670's memory runs at 1502 MHz (6 Gbps effective) for 192.3 GB/s. That is a 16.7% bandwidth advantage for the GTX 770.

Clock speeds follow the same pattern. The GTX 770 has a base clock of 1046 MHz and a boost clock of 1085 MHz. The GTX 670 runs at 915 MHz base and 980 MHz boost. The GTX 770's clocks are roughly 14% higher at base and 10.7% higher at boost. Combine the extra shaders, higher clocks, and faster memory, and the GTX 770's compute throughput rises to 3.333 TFLOPS FP32 versus 2.634 TFLOPS for the GTX 670, a 26.5% advantage. Pixel rate is 34.72 GPixel/s versus 27.44 GPixel/s, and texture rate is 138.9 GTexel/s versus 109.8 GTexel/s.

Power draw scales accordingly. The GTX 770 has a TDP of 230 W, requires a 550 W suggested power supply, and uses one 6-pin and one 8-pin power connector. The GTX 670 draws 170 W, works with a 450 W PSU, and uses two 6-pin connectors. The GTX 770 is also physically longer at 267 mm (10.5 inches) versus 241 mm (9.5 inches) for the GTX 670. Both are dual-slot cards with the same height (111 mm) and width (38 mm).

The display outputs are identical: 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2. Both support PCIe 3.0 x16, DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. Neither card has ray tracing or tensor cores, which is expected for Kepler-era hardware.

Where Each One Wins

The GTX 770 wins everywhere the database measures. If you prioritize Metal performance, the GTX 770's 26.2% lead over the GTX 670 makes it the clear choice. Vulkan workloads also favor the GTX 770 heavily at 23.2% ahead. OpenCL is the only area where the two are close, and even there the GTX 770 holds a 2.4% edge.

For gaming, the Vulkan result is the most relevant. A 23.2% advantage in Vulkan translates to a meaningful frame rate difference in most modern titles, especially at higher resolutions where memory bandwidth and shader throughput matter more. The GTX 770's extra bandwidth (224.4 GB/s versus 192.3 GB/s) and higher texture rate (138.9 GTexel/s versus 109.8 GTexel/s) should help in texture-heavy scenes.

The GTX 670 is not without merit. Its lower 170 W TDP and 450 W PSU requirement make it easier to slot into an existing system without upgrading the power supply. It is also shorter, at 241 mm versus 267 mm, which helps in compact cases. If you are building a small form factor machine or reusing an older PSU, the GTX 670 is the more practical physical fit.

In pure compute terms, the OpenCL scores show the GTX 670 can still keep pace in certain workloads. A 2.4% difference is within run-to-run variance for many applications, so the GTX 670 remains viable for OpenCL-based tasks like some rendering and video encoding pipelines.

FAQ

Q: Which card is faster in Vulkan?

A: The GTX 770. It scores 19166 in Geekbench Vulkan versus 15553 for the GTX 670, a 23.2% advantage.

Q: How much better is the GTX 770 in Metal?

A: The GTX 770 scores 9369 versus 7424 for the GTX 670, a 26.2% lead. That is the largest single benchmark gap between the two cards.

Q: Are the two cards close in any benchmark?

A: Yes, in OpenCL. The GTX 770 scores 15707 and the GTX 670 scores 15341, a difference of only 2.4%.

Q: Do both cards have the same amount of memory?

A: Yes, both have 2 GB of GDDR5 on a 256-bit bus. The GTX 770 runs it at 1753 MHz for 224.4 GB/s, while the GTX 670 runs at 1502 MHz for 192.3 GB/s.

Q: What power supply do they need?

A: The GTX 770 has a 230 W TDP and requires a 550 W PSU with one 6-pin and one 8-pin connector. The GTX 670 has a 170 W TDP and works with a 450 W PSU using two 6-pin connectors.

Q: Which card is smaller?

A: The GTX 670 is 241 mm long versus 267 mm for the GTX 770. Both are dual-slot, 111 mm tall, and 38 mm wide.

Specification Differences

| Specification | GTX 770 | GTX 670 |

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

| Generation | GeForce 700 | GeForce 600 |

| Base clock | 1046 MHz | 915 MHz |

| Boost clock | 1085 MHz | 980 MHz |

| Memory clock | 1753 MHz (7 Gbps effective) | 1502 MHz (6 Gbps effective) |

| Memory bandwidth | 224.4 GB/s | 192.3 GB/s |

| Shading units | 1536 | 1344 |

| TMUs | 128 | 112 |

| Pixel rate | 34.72 GPixel/s | 27.44 GPixel/s |

| Texture rate | 138.9 GTexel/s | 109.8 GTexel/s |

| FP32 | 3.333 TFLOPS | 2.634 TFLOPS |

| TDP | 230 W | 170 W |

| Power connectors | 1x 6-pin + 1x 8-pin | 2x 6-pin |

| Suggested PSU | 550 W | 450 W |

| Length | 267 mm (10.5 inches) | 241 mm (9.5 inches) |

| Release date | 2013-05-29 | 2012-05-09 |

| Predecessor | GeForce 600 | GeForce 500 |

| Successor | GeForce 900 | GeForce 700 |

| Launch MSRP | 399 USD | 399 USD |

The two cards share the same chip, process node, transistor count, die size, memory capacity, bus width, ROP count, slot width, height, width, bus interface, display outputs, and API support. They differ in every performance-relevant specification listed above.

The Verdict

The data is unambiguous. The GTX 770 outperforms the GTX 670 in every benchmark recorded, with the largest gaps in Metal (26.2%) and Vulkan (23.2%), and a smaller but still positive margin in OpenCL (2.4%). If you are choosing between these two for a build or an upgrade, the GTX 770 is the stronger performer in every measurable way.

The GTX 670 is not a bad card. Its OpenCL score is close to the GTX 770, and its lower power draw (170 W versus 230 W) and shorter length make it easier to install in constrained systems. It also launched at the same 399 USD MSRP as the GTX 770, so there is no price-based reason to prefer it.

For anyone who prioritizes raw performance, especially in Vulkan-based games or Metal-based applications, the GTX 770 is the only sensible pick. The extra shading units, higher clocks, faster memory, and greater bandwidth all contribute to a card that sits 15% higher in the database's average benchmark score (14747 versus 12773) and 5 percentile points higher overall (57th versus 52nd).

The GTX 670's only real advantages are practical: it fits in shorter cases and works with less power supply headroom. If those constraints are non-negotiable, the GTX 670 remains a capable choice. Otherwise, the GTX 770 is the clear winner, and the benchmark data makes that conclusion straightforward.