NVIDIA GeForce GTX 1650 vs NVIDIA GeForce GTX 1660 Ti Comparison

The NVIDIA GeForce GTX 1660 Ti is the clear performance winner over the GTX 1650 in nearly every measurable benchmark, winning 9 of 10 head-to-head tests with margins ranging from 43.3% to 247.9%. The GTX 1650 counters with a single decisive victory in Geekbench Vulkan, where it scores 33,042 against the 1660 Ti’s 5,827, an 82.4% advantage that points to a specific software optimization rather than general hardware superiority. For users prioritizing raw rasterization, compute, and DirectX performance, the GTX 1660 Ti is the only logical choice; for those targeting Vulkan-specific workloads, the GTX 1650 offers an unexpected edge. The 1660 Ti’s average benchmark score of 7,723 sits just 0.7% above the AMD Radeon 540, while the 1650’s 7,472 average places it 0.3% ahead of the AMD Radeon HD 8850M, indicating both cards are tightly clustered with low-end rivals despite their internal gap.

The Verdict

The data is unambiguous: the GTX 1660 Ti outperforms the GTX 1650 in 9 of 10 benchmarks, with the largest win coming in 3DMark Steel Nomad DX12 where it scores 1,061 versus 305, a 247.9% advantage. In PassMark G3D, the 1660 Ti scores 12,871 against 7,880, a 63.3% lead that reflects its substantially higher shader count and memory bandwidth. The 1650’s only win is Geekbench Vulkan (33,042 vs 5,827), a result so anomalous that it suggests driver or API-specific behavior rather than a general capability edge. For any user whose workload involves DirectX 9, 10, 11, or 12, OpenGL-compatible applications, or general compute, the 1660 Ti is the definitive pick. The 1650 is only relevant if the target software relies heavily on Vulkan and is insensitive to the 1660 Ti’s other advantages. Both cards are end-of-life, so this comparison is about used or legacy purchases, and the 1660 Ti’s 41st percentile versus the 1650’s 40th percentile among all GPUs confirms the former sits slightly higher in the overall hierarchy.

Where Each One Wins

The GTX 1660 Ti wins decisively across every DirectX generation tested. In PassMark DirectX 9, it scores 190 vs 124 (53.2% ahead); in DirectX 10, 66 vs 39 (69.2%); in DirectX 11, 103 vs 58 (77.6%); and in DirectX 12, 51 vs 35 (45.7%). Its compute advantage is equally clear: PassMark GPU Compute shows 5,226 vs 3,048, a 71.5% margin, and Geekbench OpenCL shows 51,029 vs 29,629 (72.2%). The 2D performance also favors the 1660 Ti at 804 vs 561 in PassMark G2D (43.3%). The GTX 1650’s sole territory is Geekbench Vulkan, where its 33,042 score crushes the 1660 Ti’s 5,827. This 82.4% swing is the largest relative difference in either direction across all tests, but it is isolated. In 3DMark Steel Nomad DX12, the 1660 Ti’s 247.9% lead is the largest absolute gap, meaning the 1650’s Vulkan win does not translate to any other API. For legacy DirectX gaming, modern DX12 titles, or OpenCL compute, the 1660 Ti is uniformly superior; for Vulkan-specific engines, the 1650 demonstrates a unique strength.

Architecture Differences

Both cards use NVIDIA’s Turing architecture on TSMC’s 12 nm process, but they diverge significantly in chip design and memory configuration. The 1660 Ti uses the TU116 chip with 6,600 million transistors on a 284 mm² die, while the 1650 uses the TU117 with 4,700 million transistors on a 200 mm² die. Transistor density is nearly identical (23.2M/mm² vs 23.5M/mm²), indicating the same process node, but the 1660 Ti packs 45% more transistors. The 1660 Ti has 1,536 shading units, 96 texture mapping units, and 48 ROPs, versus the 1650’s 896 shading units, 56 TMUs, and 32 ROPs. Neither card has ray tracing or tensor cores. Memory is another major divider: the 1660 Ti offers 6 GB of GDDR6 on a 192-bit bus with 288.0 GB/s bandwidth, while the 1650 has 4 GB of GDDR5 on a 128-bit bus with 128.1 GB/s bandwidth. The 1660 Ti’s memory clock is 1500 MHz (12 Gbps effective) versus 2001 MHz (8 Gbps effective) for the 1650, but the wider bus more than compensates. Pixel rate is 84.96 GPixel/s vs 53.28 GPixel/s, and texture rate is 169.9 GTexel/s vs 93.24 GTexel/s. FP32 compute is 5.437 TFLOPS vs 2.984 TFLOPS, with FP16 at 10.87 TFLOPS vs 5.967 TFLOPS (both 2:1). The 1660 Ti also consumes more power (120 W TDP vs 75 W) and requires a 1x 8-pin connector, while the 1650 has no power connector. Both have identical display outputs (1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a) and API support (DirectX 12_1, OpenGL 4.6, Vulkan 1.4).

FAQ

Q: Which card has higher raw compute performance?

A: The GTX 1660 Ti delivers 5.437 TFLOPS FP32 versus the GTX 1650’s 2.984 TFLOPS, an 82.2% advantage that aligns with its 71.5% lead in PassMark GPU Compute.

Q: Why does the GTX 1650 win in Geekbench Vulkan?

A: The GTX 1650 scores 33,042 in Geekbench Vulkan versus the 1660 Ti’s 5,827, an 82.4% margin. This is the only benchmark where the 1650 wins, suggesting Vulkan-specific driver optimization rather than general hardware superiority.

Q: How do memory specs differ?

A: The GTX 1660 Ti has 6 GB GDDR6 on a 192-bit bus with 288.0 GB/s bandwidth, while the GTX 1650 has 4 GB GDDR5 on a 128-bit bus with 128.1 GB/s bandwidth. The 1660 Ti’s bandwidth is 124.9% higher.

Q: What are the power requirements?

A: The GTX 1660 Ti has a 120 W TDP and requires a 1x 8-pin power connector with a suggested 300 W PSU. The GTX 1650 has a 75 W TDP, needs no power connector, and suggests a 250 W PSU.

Q: Are both cards in the same performance percentile?

A: Nearly. The GTX 1660 Ti is in the 41st percentile of all GPUs, while the GTX 1650 is in the 40th. Despite the 1660 Ti’s large internal lead, both sit near the bottom of the overall performance distribution.

Q: Which card has better DirectX 12 performance?

A: The GTX 1660 Ti scores 51 in PassMark DirectX 12 versus the 1650’s 35, a 45.7% lead. In 3DMark Steel Nomad DX12, the 1660 Ti’s 1,061 score is 247.9% higher than the 1650’s 305.

Head-to-Head Benchmarks

The GTX 1660 Ti dominates the head-to-head results with nine wins out of ten tests. The most extreme margin is in 3DMark Steel Nomad DX12, where the 1660 Ti scores 1,061 against 305, a 247.9% delta — the single largest performance gap in the entire comparison. This benchmark highlights the 1660 Ti’s architectural advantage in modern DX12 workloads, driven by its 1,536 shading units versus 896 and its 288.0 GB/s memory bandwidth versus 128.1 GB/s. In PassMark G3D, the 1660 Ti posts 12,871 versus 7,880, a 63.3% lead that corroborates the 3DMark result. The compute tests tell a similar story: Geekbench OpenCL shows 51,029 vs 29,629 (72.2%), and PassMark GPU Compute shows 5,226 vs 3,048 (71.5%). The 1660 Ti also leads in every legacy DirectX test — DirectX 9 (190 vs 124, 53.2%), DirectX 10 (66 vs 39, 69.2%), DirectX 11 (103 vs 58, 77.6%), and DirectX 12 (51 vs 35, 45.7%). Even 2D performance favors the 1660 Ti, with PassMark G2D at 804 vs 561 (43.3%). The single exception is Geekbench Vulkan, where the 1650’s 33,042 score overwhelms the 1660 Ti’s 5,827, an 82.4% reversal. This Vulkan result is the only case where the 1650’s smaller die (200 mm² vs 284 mm²) and lower transistor count (4,700M vs 6,600M) do not correlate with lower performance. Across all other tests, the 1660 Ti’s higher FP32 throughput, wider memory bus, and larger ROP count translate into consistent, often massive, wins.

Specification Differences

The two cards differ in every key specification except for architecture, process node, foundry, display outputs, and API support. The GTX 1660 Ti uses the TU116 chip (6,600 million transistors, 284 mm²), while the GTX 1650 uses the TU117 (4,700 million transistors, 200 mm²). Both are Turing on 12 nm at TSMC, with transistor densities of 23.2M/mm² and 23.5M/mm² respectively. Clock speeds differ modestly: the 1660 Ti runs at 1500 MHz base and 1770 MHz boost, while the 1650 runs at 1485 MHz base and 1665 MHz boost. The memory subsystem is a major divider: the 1660 Ti has 6 GB GDDR6 with a 192-bit bus and 288.0 GB/s bandwidth, versus the 1650’s 4 GB GDDR5 with a 128-bit bus and 128.1 GB/s bandwidth. Effective memory speed is 12 Gbps for the 1660 Ti and 8 Gbps for the 1650. Core counts scale accordingly: 1,536 shading units, 96 TMUs, and 48 ROPs on the 1660 Ti versus 896 shading units, 56 TMUs, and 32 ROPs on the 1650. Pixel rate is 84.96 GPixel/s vs 53.28 GPixel/s, and texture rate is 169.9 GTexel/s vs 93.24 GTexel/s. FP32 compute is 5.437 TFLOPS vs 2.984 TFLOPS, with FP16 at 10.87 TFLOPS vs 5.967 TFLOPS (both 2:1). Power consumption differs substantially: 120 W TDP with a 1x 8-pin connector and 300 W suggested PSU for the 1660 Ti, versus 75 W TDP with no connector and 250 W suggested PSU for the 1650. Neither has ray tracing or tensor cores. Both cards are dual-slot, 229 mm long, 111 mm tall, and 35 mm wide, with identical display outputs (1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a) and API support (DirectX 12_1, OpenGL 4.6, Vulkan 1.4).