NVIDIA GeForce GTX 1650 vs NVIDIA GeForce GTX 970 Comparison

Head-to-Head Benchmarks

The benchmark data presents a decisive picture: the NVIDIA GeForce GTX 970 wins 9 of the 10 shared tests, while the GTX 1650 takes a single, but significant, victory. The most striking result is in Geekbench Vulkan, where the GTX 1650 scores 33042 against the GTX 970’s 20005, a 65.2% margin. This is the largest delta in either direction across the entire comparison, indicating a fundamental advantage in Vulkan workloads for the newer architecture.

Beyond that outlier, the GTX 970 consistently leads by double-digit percentages. In the 3DMark Steel Nomad DX12 test, the GTX 970 scores 369 against 305 for the GTX 1650, a 17.3% advantage. The Passmark suite reinforces this trend: the GTX 970 is 18.3% ahead in DirectX 11 (71 vs 58), 14.6% ahead in DirectX 12 (41 vs 35), and 13.3% ahead in DirectX 9 (143 vs 124). The gap widens further in compute-oriented tasks. Passmark GPU Compute shows the GTX 970 at 4073, which is 25.2% higher than the GTX 1650’s 3048. Even in 2D performance, where architecture differences often matter less, the GTX 970 leads by 26.6% (764 vs 561).

The overall average benchmark scores confirm the pattern, though the margin narrows. The GTX 1650 posts an average benchmark score of 7472, while the GTX 970 averages 7157. This apparent contradiction — the GTX 1650 having a higher average despite losing nearly every head-to-head test — stems from the fact that the GTX 970 lacks a Geekbench Metal score in the shared test set, while the GTX 1650 has no such missing data point. When directly compared on identical tests, the GTX 970’s superiority is unambiguous.

Architecture Differences

The two cards belong to different architectural generations. The GTX 1650 is built on the Turing architecture (chip TU117) using a 12 nm process at TSMC, while the GTX 970 uses the Maxwell 2.0 architecture (chip GM204) on a 28 nm process, also at TSMC. This process shrink is substantial: the GTX 1650 packs 4,700 million transistors into a 200 mm² die, yielding a density of 23.5M transistors per mm². The GTX 970, by contrast, has 5,200 million transistors spread across a much larger 398 mm² die, resulting in only 13.1M transistors per mm². The newer card is therefore far more efficient in terms of transistor density, which directly contributes to its dramatically lower power draw.

The compute resources differ sharply. The GTX 970 has 1664 shading units, 104 texture mapping units (TMUs), and 56 render output units (ROPs). The GTX 1650 is smaller in every category: 896 shading units, 56 TMUs, and 32 ROPs. Despite this, the GTX 1650’s higher clocks — 1485 MHz base and 1665 MHz boost versus 1050 MHz base and 1178 MHz boost for the GTX 970 — partially compensate. The pixel rate tells this story: the GTX 970 reaches 65.97 GPixel/s against 53.28 GPixel/s for the GTX 1650. The texture rate gap is smaller: 122.5 GTexel/s versus 93.24 GTexel/s. In raw FP32 throughput, the GTX 970 leads with 3.920 TFLOPS versus 2.984 TFLOPS.

Memory configuration is another differentiator. Both cards have 4 GB of GDDR5, but the GTX 970 uses a 256-bit bus width, delivering 224.4 GB/s of bandwidth, while the GTX 1650 is constrained to a 128-bit bus, yielding 128.1 GB/s. The GTX 1650’s memory runs at 2001 MHz (8 Gbps effective) compared to the GTX 970’s 1753 MHz (7 Gbps effective), but the narrower bus is the limiting factor. The GTX 1650 also supports FP16 at 5.967 TFLOPS with a 2:1 ratio, a feature the GTX 970 does not expose.

Where Each One Wins

The GTX 970 is the clear winner for traditional rasterization workloads. Its leads across DirectX 9, 10, 11, and 12 tests are consistent, ranging from 13.3% to 18.3%. The 3DMark Steel Nomad DX12 result at 17.3% ahead reinforces that this is not a legacy-only advantage — the older card holds up in modern DX12 titles. For compute-heavy tasks, the GTX 970’s 25.2% lead in Passmark GPU Compute and its 4073 score indicate it is the stronger choice for general-purpose GPU workloads.

The GTX 1650’s single win is in Geekbench Vulkan, where it scores 33042 versus 20005. This 65.2% margin is so large that it suggests the Turing architecture’s Vulkan driver implementation or hardware scheduling is markedly more efficient than Maxwell’s. For developers or users running Vulkan-based applications, this could be a decisive factor. The GTX 1650 also benefits from substantially lower power requirements: its 75 W TDP and lack of power connectors contrast sharply with the GTX 970’s 148 W TDP and dual 6-pin connectors. The newer card is also shorter at 229 mm versus 267 mm, and thinner at 35 mm versus 40 mm, making it far easier to install in compact systems.

FAQ

Q: Which card is faster in DirectX 11?

A: The GTX 970 is 18.3% faster in Passmark DirectX 11, scoring 71 versus 58 for the GTX 1650.

Q: Is the GTX 1650 better in any benchmark?

A: Yes, it wins Geekbench Vulkan by a 65.2% margin (33042 vs 20005). It does not win any other shared test.

Q: How do the power requirements compare?

A: The GTX 1650 has a 75 W TDP with no power connectors, while the GTX 970 has a 148 W TDP and requires two 6-pin connectors. The suggested PSU is 250 W for the GTX 1650 and 300 W for the GTX 970.

Q: Do both cards support the same DirectX version?

A: Yes, both support DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4.

Q: Which card has more memory bandwidth?

A: The GTX 970 has significantly more, at 224.4 GB/s, compared to 128.1 GB/s for the GTX 1650, due to its 256-bit bus versus 128-bit.

Q: What is the average benchmark score difference?

A: The GTX 1650 has a higher average benchmark score at 7472, versus 7157 for the GTX 970, though this includes non-shared tests.

The Verdict

For users prioritizing raw performance in most gaming and compute scenarios, the GTX 970 is the stronger card. Its wins across DirectX 9 through 12, 3DMark, and GPU compute are decisive, with margins ranging from 13.3% to 26.6%. The 18.2% lead in Passmark G3D (9638 vs 7880) is representative of overall gaming performance. Its 3.920 TFLOPS FP32 throughput and 224.4 GB/s bandwidth give it a clear hardware advantage that clock speed alone cannot close.

However, the GTX 1650 is not without merit. Its Vulkan performance is exceptional, beating the GTX 970 by 65.2%, which suggests it is the better choice for Vulkan-specific applications. Its 75 W TDP, lack of external power connectors, and smaller physical footprint make it ideal for low-power or small-form-factor builds. The GTX 1650 also offers FP16 support at 5.967 TFLOPS, which the GTX 970 lacks entirely.

The decision hinges on context. For a standard gaming PC with adequate power delivery and case space, the GTX 970’s higher frame rates across most APIs make it the data-backed pick. For a compact, power-constrained system or a workload centered on Vulkan, the GTX 1650’s efficiency and driver advantage are compelling. The data does not support a universal winner — it supports a situational one.

Specification Differences

| Specification | GTX 1650 | GTX 970 |

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

| Architecture | Turing | Maxwell 2.0 |

| Process Node | 12 nm | 28 nm |

| Transistors | 4,700 million | 5,200 million |

| Die Size | 200 mm² | 398 mm² |

| Transistor Density | 23.5M / mm² | 13.1M / mm² |

| Base Clock | 1485 MHz | 1050 MHz |

| Boost Clock | 1665 MHz | 1178 MHz |

| Memory Clock | 2001 MHz (8 Gbps effective) | 1753 MHz (7 Gbps effective) |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 128.1 GB/s | 224.4 GB/s |

| Shading Units | 896 | 1664 |

| TMUs | 56 | 104 |

| ROPs | 32 | 56 |

| Pixel Rate | 53.28 GPixel/s | 65.97 GPixel/s |

| Texture Rate | 93.24 GTexel/s | 122.5 GTexel/s |

| FP32 | 2.984 TFLOPS | 3.920 TFLOPS |

| FP16 | 5.967 TFLOPS (2:1) | Not specified |

| TDP | 75 W | 148 W |

| Power Connectors | None | 2x 6-pin |

| Suggested PSU | 250 W | 300 W |

| Length | 229 mm (9 inches) | 267 mm (10.5 inches) |

| Width | 35 mm (1.4 inches) | 40 mm (1.6 inches) |

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