NVIDIA GeForce GTX 1650 SUPER vs NVIDIA GeForce GTX 1660 Comparison

The Verdict

The data splits these two cards cleanly. The NVIDIA GeForce GTX 1660 is the stronger performer in nine of the ten head-to-head tests, with its biggest win coming in the 3DMark Steel Nomad DX12 test where it scores 1065 against the 1650 SUPER's 352 — a staggering 202.6% advantage. That is not a small gap; it is a generational-style leap in a modern API workload. The 1650 SUPER's only win is in Geekbench Vulkan, where it edges ahead by a narrow 0.8% (50519 vs 50137), which is within noise for most real-world use.

For the builder, the choice is straightforward: the GTX 1660 is the default pick if you want higher frame rates and more headroom in DX12-heavy titles. The 1650 SUPER makes sense only if you have a hard constraint on power draw — its 100 W TDP is 20 W lower than the 1660's 120 W — or if you specifically need that small Vulkan edge. The 1660 also carries a 6 GB memory buffer versus 4 GB on the 1650 SUPER, which matters for texture-heavy games and future-proofing. Neither card is a current-generation part; both are end-of-life, but the 1660's overall average benchmark score of 11680 versus 11047 for the 1650 SUPER tells the story: about 5.7% more aggregate performance, with the gap widening significantly in specific modern workloads.

Architecture Differences

Both cards share the same TU116 chip, Turing architecture, and 12 nm TSMC process node, with identical transistor counts of 6,600 million and die size of 284 mm². The core hardware is the same silicon, but NVIDIA configured it differently. The GTX 1660 has 1408 shading units, 88 texture mapping units, and 48 ROPs. The 1650 SUPER is cut down to 1280 shading units, 80 TMUs, and 32 ROPs. That 128-shader and 16-ROP deficit directly explains the 1660's higher pixel rate (85.68 GPixel/s vs 55.20 GPixel/s) and texture rate (157.1 GTexel/s vs 138.0 GTexel/s).

Memory is where the two diverge most sharply. The 1660 uses 6 GB of GDDR5 on a 192-bit bus, yielding 192.1 GB/s of bandwidth. The 1650 SUPER uses 4 GB of GDDR6 on a 128-bit bus, producing 192.0 GB/s — nearly identical bandwidth despite the narrower bus, thanks to faster 12 Gbps effective memory clocks versus the 1660's 8 Gbps effective. The 1650 SUPER's memory clock runs at 1500 MHz, while the 1660's runs at 2001 MHz, but the effective data rate flips that. This is a classic trade-off: more capacity on the 1660, faster memory technology on the 1650 SUPER, but with the same practical throughput.

Both cards support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, and neither has ray tracing or tensor cores. Pixel and texture throughput favor the 1660 because of its higher ROP and TMU counts. The 1650 SUPER's boost clock is 1725 MHz, 60 MHz lower than the 1660's 1785 MHz, contributing to the 1660's higher FP32 compute of 5.027 TFLOPS versus 4.416 TFLOPS on the smaller card.

Head-to-Head Benchmarks

The head-to-head results are lopsided, but the magnitude varies by workload. The biggest delta is in 3DMark Steel Nomad DX12: 1065 vs 352, a 202.6% win for the 1660. This is not a typical difference; it suggests the 1660's larger memory bus and higher ROP count are critical for modern DX12 rendering pipelines. In PassMark DirectX 10, the 1660 wins 61 to 50, a 22% margin. PassMark DirectX 9 shows a 19.6% lead (177 vs 148), and PassMark G3D has the 1660 ahead by 14.4% (11646 vs 10179).

Compute workloads also favor the 1660. PassMark GPU Compute scores 4963 vs 4477, a 10.9% advantage, which aligns with its higher FP32 throughput. Geekbench OpenCL shows a 9.1% lead (47850 vs 43875). The 1660 also wins the older API tests: PassMark DirectX 11 (79 vs 73, 8.2%) and PassMark DirectX 12 (49 vs 45, 8.9%). Even the 2D test goes to the 1660, 776 vs 749, a 3.6% margin.

The single bright spot for the 1650 SUPER is Geekbench Vulkan: 50519 vs 50137, a 0.8% win. That is a narrow victory, and one test does not reverse the overall trend. The 1660 wins nine of ten tests outright, and its losses are only in a single synthetic Vulkan workload that may not reflect typical gaming.

Specification Differences

The two cards differ in several key specification fields. The GTX 1660 has a boost clock of 1785 MHz; the 1650 SUPER boosts to 1725 MHz. The 1660's base clock is 1530 MHz, while the 1650 SUPER also lists 1530 MHz — identical base, but lower boost. Memory configuration: the 1660 has 6 GB GDDR5 on a 192-bit bus; the 1650 SUPER has 4 GB GDDR6 on a 128-bit bus. Effective memory speed is 8 Gbps on the 1660 and 12 Gbps on the 1650 SUPER, though memory clock differs (2001 MHz vs 1500 MHz). Bandwidth is effectively the same: 192.1 GB/s vs 192.0 GB/s.

Shading units: 1408 vs 1280. TMUs: 88 vs 80. ROPs: 48 vs 32. Pixel rate: 85.68 GPixel/s vs 55.20 GPixel/s. Texture rate: 157.1 GTexel/s vs 138.0 GTexel/s. FP32: 5.027 TFLOPS vs 4.416 TFLOPS. FP16: 10.05 TFLOPS vs 8.832 TFLOPS (both 2:1). TDP: 120 W vs 100 W. Power connector: 1x 8-pin on the 1660, 1x 6-pin on the 1650 SUPER. Launch MSRP: 219 USD for the 1660, 159 USD for the 1650 SUPER. Release dates: 2019-03-13 for the 1660, 2019-11-21 for the 1650 SUPER. Both are dual-slot, 229 mm long, 111 mm tall, 35 mm wide, use PCIe 3.0 x16, and have the same display outputs (1x DVI, 1x HDMI 2.0, 1x DisplayPort 1.4a). Suggested PSU is 300 W for both. Both share the same transistor density of 23.2M / mm².

FAQ

Q: Which card is faster in modern DX12 games?

A: The GTX 1660 wins decisively. In the 3DMark Steel Nomad DX12 test, it scores 1065 versus the 1650 SUPER's 352, a 202.6% advantage. That suggests the 1660's larger memory bus and higher ROP count provide a substantial edge in DX12 workloads.

Q: Is the 1650 SUPER ever the better choice?

A: Only if you prioritize lower power draw. The 1650 SUPER has a 100 W TDP versus 120 W for the 1660, and it wins Geekbench Vulkan by 0.8% (50519 vs 50137). If your PSU or cooling is tight, the 1650 SUPER is the more frugal option.

Q: Does the 1650 SUPER's GDDR6 memory make up for its smaller size?

A: No. Despite using faster GDDR6 at 12 Gbps effective versus the 1660's GDDR5 at 8 Gbps, the 1650 SUPER's 128-bit bus limits bandwidth to 192.0 GB/s, virtually identical to the 1660's 192.1 GB/s. The 1660 still has more capacity (6 GB vs 4 GB) and wins almost every benchmark.

Q: How do these cards compare to their nearest rivals?

A: The 1660's average benchmark score of 11680 is 0.5% ahead of the AMD Radeon RX 7800 XT (11627) and 1.3% ahead of the AMD Radeon Pro 5500M (11528), but 1.4% behind the AMD Radeon RX 6500 XT (11842). The 1650 SUPER's average of 11047 is 0.2% behind the AMD Radeon RX 550 (11075) and 1.5% ahead of the NVIDIA GeForce MX350 (10883).

Q: Which card has better compute performance?

A: The GTX 1660. It scores 4963 in PassMark GPU Compute versus 4477 for the 1650 SUPER, a 10.9% lead. Its higher FP32 throughput of 5.027 TFLOPS versus 4.416 TFLOPS supports this result.

Q: Are the physical dimensions the same?

A: Yes. Both cards are 229 mm long, 111 mm tall, and 35 mm wide, and both are dual-slot. The power connectors differ: the 1660 uses a 1x 8-pin, the 1650 SUPER uses a 1x 6-pin.

Where Each One Wins

The GTX 1660 wins in every meaningful gaming and compute scenario from the data. Its 202.6% lead in 3DMark Steel Nomad DX12 makes it the clear pick for modern DirectX 12 titles, and its 14.4% advantage in PassMark G3D confirms broad gaming superiority. The 1660 also wins legacy API tests — DirectX 9 by 19.6%, DirectX 10 by 22%, DirectX 11 by 8.2% — so it is not just a modern-API specialist. Compute users get a 10.9% edge in PassMark GPU Compute and a 9.1% lead in Geekbench OpenCL. The 6 GB memory capacity gives it more headroom for texture-heavy games than the 4 GB on the 1650 SUPER.

The 1650 SUPER's only win is Geekbench Vulkan by 0.8%, which is a marginal and workload-specific result. Its real advantage is efficiency: a 100 W TDP versus 120 W, and a 6-pin connector instead of an 8-pin, making it easier to fit into pre-built systems with modest PSUs. If you are building a low-power HTPC or a secondary rig with a 300 W PSU (the suggested rating for both), the 1650 SUPER's lower draw is the only practical reason to choose it. For any other use case, the data says the GTX 1660 is the superior card, with a launch MSRP of 219 USD versus 159 USD for the 1650 SUPER — the price difference buys substantially more performance in most tests.