NVIDIA GeForce GTX 1660 Ti vs NVIDIA GeForce GTX 970 Comparison

The NVIDIA GeForce GTX 1660 Ti and the NVIDIA GeForce GTX 970 are separated by nearly five years of GPU architecture, yet they often appear in the same used-market conversations. The data shows a clear generational shift: the 1660 Ti wins 9 of 10 head-to-head benchmarks, but the GTX 970 still holds a surprising lead in one key API test. This comparison breaks down where the Turing-based card decisively outperforms the Maxwell veteran, and where the older card’s legacy silicon remains relevant.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The NVIDIA GeForce GTX 1660 Ti has an average benchmark score of 7723, compared to the GTX 970’s 7157. That is a difference of 566 points, placing the 1660 Ti in the 41st percentile of all GPUs while the GTX 970 sits in the 39th percentile.

Q: How large is the performance gap in the most demanding DX12 test?

A: In the 3DMark Steel Nomad DX12 test, the GTX 1660 Ti scores 1061 versus the GTX 970’s 369. That is a 187.5% delta, making it the single largest margin of victory for the 1660 Ti across all head-to-head benchmarks.

Q: Does the older GTX 970 win any benchmark against the 1660 Ti?

A: Yes. In the Geekbench Vulkan test, the GTX 970 scores 20005 while the 1660 Ti scores only 5827. This represents a -70.9% delta for the 1660 Ti, meaning the GTX 970 is roughly 3.4 times faster in that specific workload.

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

A: The GTX 1660 Ti features 6 GB of GDDR6 memory on a 192-bit bus, delivering 288.0 GB/s of bandwidth. The GTX 970 has 4 GB of GDDR5 on a 256-bit bus, providing 224.4 GB/s. The 1660 Ti offers 50% more memory capacity and 28.3% more bandwidth.

Q: How do the two cards compare in compute performance?

A: In the Passmark GPU Compute test, the 1660 Ti scores 5226 against the GTX 970’s 4073, a 28.3% advantage. The 1660 Ti also delivers 5.437 TFLOPS of FP32 performance versus the GTX 970’s 3.920 TFLOPS.

Q: Which card is more power-efficient on paper?

A: The GTX 1660 Ti has a TDP of 120 W, while the GTX 970 is rated at 148 W. Both cards recommend a 300 W power supply, but the 1660 Ti achieves higher performance with a 28 W lower thermal design power.

Architecture Differences

The two cards represent entirely different architectural generations. The GTX 1660 Ti uses the TU116 chip built on Turing architecture, fabricated on TSMC’s 12 nm process. It packs 6,600 million transistors into a 284 mm² die, yielding a transistor density of 23.2M per mm². The GTX 970, by contrast, uses the GM204 chip on Maxwell 2.0 architecture, built on a 28 nm process. It contains 5,200 million transistors on a larger 398 mm² die, resulting in a lower density of 13.1M per mm².

The shader configuration differs notably. The GTX 970 actually has more shading units (1664) and TMUs (104) than the 1660 Ti (1536 shading units, 96 TMUs). However, the 1660 Ti has fewer ROPs (48 versus 56). Despite having less raw hardware, the 1660 Ti achieves higher clock speeds: 1500 MHz base and 1770 MHz boost, compared to the GTX 970’s 1050 MHz base and 1178 MHz boost. That 592 MHz boost clock advantage helps explain the performance gap.

Memory architecture is another major split. The 1660 Ti uses 6 GB of GDDR6 memory at 1500 MHz (12 Gbps effective), while the GTX 970 uses 4 GB of GDDR5 at 1753 MHz (7 Gbps effective). The newer memory type gives the 1660 Ti a bandwidth advantage of 288.0 GB/s versus 224.4 GB/s, even though the GTX 970 has a wider 256-bit bus compared to 192-bit.

The GTX 970 lacks FP16 support entirely, while the 1660 Ti delivers 10.87 TFLOPS of FP16 performance at a 2:1 ratio. Both cards support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, so API compatibility is not a differentiator. The 1660 Ti uses a single 8-pin power connector, while the GTX 970 requires two 6-pin connectors. The 1660 Ti is also shorter at 229 mm versus the GTX 970’s 267 mm, and slightly thinner at 35 mm versus 40 mm.

Head-to-Head Benchmarks

The 3DMark Steel Nomad DX12 test delivers the most dramatic result. The GTX 1660 Ti scores 1061 against the GTX 970’s 369, a 187.5% delta. This is not a marginal win; the newer card is nearly three times faster in this modern DX12 workload.

In Geekbench OpenCL, the 1660 Ti scores 51029 versus 29982, a 70.2% advantage. This reflects the compute-heavy nature of the test, where the 1660 Ti’s higher FP32 throughput (5.437 TFLOPS) and faster GDDR6 memory pay off. The Passmark G3D test shows a 33.5% lead for the 1660 Ti (12871 versus 9638), which aligns with the overall 7723 versus 7157 average score gap.

The legacy DirectX tests favor the newer card consistently. In Passmark DirectX 11, the 1660 Ti scores 103 versus 71, a 45.1% delta. DirectX 10 shows 66 versus 46, a 43.5% delta. Even DirectX 9, a test that should favor older architectures, goes to the 1660 Ti at 190 versus 143, a 32.9% delta. The smallest 3D win is in Passmark DirectX 12, where the 1660 Ti scores 51 versus 41, a 24.4% delta.

Compute workloads also favor the 1660 Ti. Passmark GPU Compute shows 5226 versus 4073, a 28.3% delta. The 2D test is close: 804 versus 764, only a 5.2% delta, suggesting that desktop and basic 2D workloads are not a meaningful differentiator.

The single exception is Geekbench Vulkan. Here, the GTX 970 scores 20005 while the 1660 Ti manages only 5827. This is a -70.9% delta for the 1660 Ti, meaning the older card is substantially faster in this specific Vulkan implementation. This result is anomalous given the 1660 Ti’s wins everywhere else, but it is the only benchmark where the GTX 970 takes the victory.

The Verdict

The data is unambiguous: the NVIDIA GeForce GTX 1660 Ti is the faster card in almost every measurable way. It wins 9 of 10 head-to-head benchmarks, with an average score advantage of 566 points. Its 187.5% lead in 3DMark Steel Nomad DX12 indicates that modern, API-heavy workloads will run dramatically better on the 1660 Ti. The card also offers more memory (6 GB versus 4 GB), higher bandwidth (288.0 GB/s versus 224.4 GB/s), and lower power consumption (120 W versus 148 W).

The GTX 970’s lone Vulkan win is notable but narrow in scope. A -70.9% delta means the 1660 Ti is far behind in that one test, but it does not offset the 70.2% OpenCL advantage or the 33.5% G3D lead for the 1660 Ti. For a buyer choosing between these two today, the 1660 Ti is the logical pick for gaming and general compute. Its 12 nm Turing architecture delivers modern features and efficiency that the 28 nm Maxwell chip cannot match.

The GTX 970 remains a functional card for lighter workloads, but its 39th percentile ranking and lower scores across most tests make it a legacy option. The 1660 Ti’s 41st percentile ranking is not a huge step up in relative terms, but the absolute performance differences and memory advantages are decisive. Choose the 1660 Ti unless you have a specific Vulkan workload that exploits the GTX 970’s unexpected strength.

Specification Differences

| Specification | NVIDIA GeForce GTX 1660 Ti | NVIDIA GeForce GTX 970 |

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

| Architecture | Turing | Maxwell 2.0 |

| Process Node | 12 nm | 28 nm |

| Transistors | 6,600 million | 5,200 million |

| Die Size | 284 mm² | 398 mm² |

| Transistor Density | 23.2M / mm² | 13.1M / mm² |

| Base Clock | 1500 MHz | 1050 MHz |

| Boost Clock | 1770 MHz | 1178 MHz |

| Memory Size | 6 GB | 4 GB |

| Memory Type | GDDR6 | GDDR5 |

| Memory Bus | 192 bit | 256 bit |

| Memory Bandwidth | 288.0 GB/s | 224.4 GB/s |

| Shading Units | 1536 | 1664 |

| TMUs | 96 | 104 |

| ROPs | 48 | 56 |

| FP32 Performance | 5.437 TFLOPS | 3.920 TFLOPS |

| FP16 Performance | 10.87 TFLOPS (2:1) | null |

| Pixel Rate | 84.96 GPixel/s | 65.97 GPixel/s |

| Texture Rate | 169.9 GTexel/s | 122.5 GTexel/s |

| TDP | 120 W | 148 W |

| Power Connectors | 1x 8-pin | 2x 6-pin |

| Length | 229 mm | 267 mm |

| Width | 35 mm | 40 mm |

| Display Outputs | 1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2 |

| Release Date | 2019-02-21 | 2014-09-18 |

| Launch MSRP | 279 USD | 329 USD |

Where Each One Wins

The GTX 1660 Ti wins in every category that matters for modern gaming and compute. Its 187.5% lead in 3DMark Steel Nomad DX12 makes it the clear choice for DX12 titles. The 70.2% OpenCL advantage covers GPU-accelerated applications that rely on OpenCL for compute tasks. For DirectX 9, 10, 11, and 12 legacy tests, the 1660 Ti is consistently 24.4% to 45.1% faster. The 33.5% G3D win and 28.3% compute win round out a comprehensive sweep of 3D and general-purpose workloads. Its 6 GB memory buffer and 288.0 GB/s bandwidth give it headroom for higher-resolution textures and modern game assets.

The GTX 970 wins exactly one benchmark: Geekbench Vulkan, with a score of 20005 versus 5827. This makes it the better pick for workloads that are heavily optimized for Vulkan and happen to hit the GTX 970’s architectural strengths. The 2D test is nearly a tie (764 versus 804), so neither card has a meaningful edge in desktop use. The GTX 970 also holds advantages in raw hardware counts—1664 shading units, 104 TMUs, and 56 ROPs—but those do not translate into wins outside of the Vulkan test.

For a practical builder, the choice is simple. The GTX 1660 Ti is faster in 9 of 10 benchmarks, uses less power, runs shorter and cooler, and offers more memory. The GTX 970 is only relevant if you have a specific Vulkan-centric workload or need the extra DisplayPort outputs (3 versus 1). Otherwise, the 1660 Ti is the superior card by every measurable benchmark metric.