NVIDIA GeForce GTX 650 Ti vs NVIDIA GeForce GTX 970 Comparison

The Verdict

The benchmark database presents a decisive outcome: the NVIDIA GeForce GTX 970 wins every recorded head-to-head comparison against the GeForce GTX 650 Ti. In the two shared tests, the GTX 970 leads by substantial margins, 73.7% in Geekbench OpenCL and 58.9% in Geekbench Vulkan. The GTX 650 Ti does not win a single recorded benchmark. For any user choosing between these two cards based on measured performance, the GTX 970 is the clear selection.

However, the database also reveals a nuance that complicates the simple verdict. The GTX 970's average benchmark score across all its recorded tests is 7,157, which places it in the 39th percentile of all GPUs. The GTX 650 Ti, with only two recorded benchmarks, averages 8,053 and sits in the 42nd percentile. This discrepancy stems from the different test sets applied to each card. The GTX 970 was evaluated across eleven tests, including DirectX 9 through 12, compute workloads, and 2D graphics, while the GTX 650 Ti was only tested in OpenCL and Vulkan. The GTX 650 Ti's percentile reflects its performance in those two compute-oriented tests only.

For buyers, the guidance is straightforward: the GTX 970 is the superior card in raw compute and graphics workloads measured by the database. The GTX 650 Ti, despite its higher percentile ranking, achieves that position within a narrower test scope. Users needing modern API support, higher memory capacity, or stronger compute performance should select the GTX 970. Users constrained by power or physical space might consider the GTX 650 Ti, but they would sacrifice significant performance.

Where Each One Wins

The database records zero benchmark wins for the GTX 650 Ti. In the two tests where both cards appear, the GTX 970 wins both. The GTX 650 Ti's recorded scores are 7,877 in Geekbench OpenCL and 8,229 in Geekbench Vulkan. The GTX 970 scores 29,982 and 20,005 in the same respective tests.

The GTX 970 also demonstrates versatility across a broader benchmark suite. Its Passmark results show 143 in DirectX 9, 71 in DirectX 11, 46 in DirectX 10, and 41 in DirectX 12. The compute score is 4,073, the 3D score is 9,638, and the 2D score is 764. The GTX 970 additionally records 13,595 in Geekbench Metal and 369 in 3DMark Steel Nomad DX12. None of these tests were applied to the GTX 650 Ti, so no direct comparison exists for those workloads.

The GTX 650 Ti's strengths, if any, would have to be inferred from its architectural characteristics rather than measured victories. It consumes less power at 110 W versus 148 W, uses a single slot versus dual slots, and requires a 1x 6-pin power connector versus 2x 6-pin. Its smaller physical footprint, 145 mm length versus 267 mm, makes it suitable for compact systems. But in terms of benchmark performance, the GTX 650 Ti has no recorded advantage.

Architecture Differences

The two cards come from different NVIDIA generations and architectures. The GTX 650 Ti uses the GK106S chip based on the Kepler architecture, part of the GeForce 600 generation. The GTX 970 uses the GM204 chip based on Maxwell 2.0, part of the GeForce 900 generation. Both are manufactured by TSMC on a 28 nm process, but the transistor counts differ substantially: the GTX 650 Ti packs 2,540 million transistors on a 221 mm² die, while the GTX 970 contains 5,200 million transistors on a 398 mm² die. The transistor density is higher on the GTX 970 at 13.1 million per mm² versus 11.5 million per mm².

The compute resources scale accordingly. The GTX 650 Ti has 768 shading units, 64 texture mapping units, and 16 render output units. The GTX 970 has 1,664 shading units, 104 TMUs, and 56 ROPs. These are more than double, nearly double, and more than triple the respective counts. The pixel rate for the GTX 650 Ti is 14.85 GPixel/s versus 65.97 GPixel/s for the GTX 970. Texture rate is 59.39 GTexel/s versus 122.5 GTexel/s. Floating point performance is 1,425.4 GFLOPS versus 3.920 TFLOPS.

Memory configurations also differ. The GTX 650 Ti has 1,024 MB of GDDR5 on a 128-bit bus with 86.40 GB/s bandwidth. The GTX 970 has 4 GB of GDDR5 on a 256-bit bus with 224.4 GB/s bandwidth. Memory clocks are 1,350 MHz (5.4 Gbps effective) for the older card versus 1,753 MHz (7 Gbps effective) for the newer one. The GTX 970 also has defined base and boost clocks at 1,050 MHz and 1,178 MHz, while the GTX 650 Ti's core clocks are not recorded in the database.

API support differs. The GTX 650 Ti supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The GTX 970 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4. The newer card also supports Metal as shown by its Geekbench Metal score. Display outputs differ: the GTX 650 Ti has 2x DVI and 1x mini-HDMI 1.4a, while the GTX 970 has 1x DVI, 1x HDMI 2.0, and 3x DisplayPort 1.2.

FAQ

Q: Which card has better compute performance in the database?

A: The GTX 970 scores 29,982 in Geekbench OpenCL versus 7,877 for the GTX 650 Ti, a 73.7% advantage. In Geekbench Vulkan, the GTX 970 scores 20,005 versus 8,229, a 58.9% advantage.

Q: Does the older GTX 650 Ti support newer APIs?

A: The GTX 650 Ti supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. The GTX 970 supports DirectX 12 (12_1), OpenGL 4.6, and Vulkan 1.4, which is a newer Vulkan version.

Q: How do the memory capacities compare?

A: The GTX 650 Ti has 1,024 MB of GDDR5 memory on a 128-bit bus. The GTX 970 has 4 GB of GDDR5 memory on a 256-bit bus. Bandwidth is 86.40 GB/s versus 224.4 GB/s.

Q: Which card requires less power and space?

A: The GTX 650 Ti has a 110 W TDP, uses a single slot, and measures 145 mm in length. The GTX 970 has a 148 W TDP, uses dual slots, and measures 267 mm in length. Both require a 300 W suggested power supply.

Q: Are both cards still in production?

A: No. The database lists both as end-of-life production status. The GTX 650 Ti was released on 2012-10-08, and the GTX 970 was released on 2014-09-18.

Q: What is the GTX 970's performance spread across its benchmark suite?

A: The GTX 970's Passmark scores range from 41 in DirectX 12 to 143 in DirectX 9. Its G3D score is 9,638, GPU compute is 4,073, and G2D is 764. Geekbench Metal is 13,595, and 3DMark Steel Nomad DX12 is 369.

Head-to-Head Benchmarks

The database includes exactly two head-to-head comparisons between these cards. Both are Geekbench tests, and both show the GTX 970 winning by wide margins.

In Geekbench OpenCL, the GTX 650 Ti scores 7,877. The GTX 970 scores 29,982. The percentage difference is 73.7% in favor of the GTX 970. This means the GTX 970 delivers roughly four times the OpenCL compute performance. For context, the GTX 650 Ti's nearest rivals in the database are the GTX 880M at 8,040 (0.2% ahead), the GTX 650 Ti Boost at 8,067 (0.2% behind), the Quadro P5000 at 8,039 (0.2% ahead), and the GRID K2 at 8,080 (0.3% behind). The GTX 970's nearest rivals are the Intel Iris Pro Graphics P580 at 7,170 (0.2% behind), the GTX 560 SE at 7,171 (0.2% behind), the Radeon Vega 8 Mobile at 7,203 (0.6% behind), and the GTX 750 at 7,222 (0.9% behind). The GTX 970's OpenCL score of 29,982 is not directly comparable to these rival averages because those averages span multiple tests, but the magnitude of the GTX 970's OpenCL result far exceeds its own average benchmark score of 7,157.

In Geekbench Vulkan, the GTX 650 Ti scores 8,229. The GTX 970 scores 20,005. The percentage difference is 58.9% in favor of the GTX 970. The GTX 970 more than doubles the Vulkan performance of the older card. Notably, the GTX 650 Ti's Vulkan score of 8,229 is slightly higher than its OpenCL score of 7,877, suggesting the Kepler architecture handles Vulkan workloads reasonably well relative to its own OpenCL performance. The GTX 970's Vulkan score of 20,005 is lower than its OpenCL score of 29,982, indicating that Maxwell 2.0 excels more in OpenCL than in Vulkan, though it remains far ahead of the GTX 650 Ti in both.

The GTX 970 also has scores in tests that were not applied to the GTX 650 Ti. These include Geekbench Metal at 13,595, 3DMark Steel Nomad DX12 at 369, Passmark DirectX 9 at 143, DirectX 10 at 46, DirectX 11 at 71, DirectX 12 at 41, G2D at 764, G3D at 9,638, and GPU compute at 4,073. Without corresponding GTX 650 Ti scores, these results cannot be compared directly, but they demonstrate the breadth of testing applied to the newer card.

Specification Differences

The following table summarizes only the fields where the two cards differ in the database.

| Specification | GTX 650 Ti | GTX 970 |

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

| Chip | GK106S | GM204 |

| Architecture | Kepler | Maxwell 2.0 |

| Generation | GeForce 600 | GeForce 900 |

| Transistors | 2,540 million | 5,200 million |

| Die Size | 221 mm² | 398 mm² |

| Transistor Density | 11.5M / mm² | 13.1M / mm² |

| Base Clock | Not recorded | 1050 MHz |

| Boost Clock | Not recorded | 1178 MHz |

| Memory Clock | 1350 MHz, 5.4 Gbps effective | 1753 MHz, 7 Gbps effective |

| Memory Size | 1024 MB | 4 GB |

| Memory Bus Width | 128 bit | 256 bit |

| Memory Bandwidth | 86.40 GB/s | 224.4 GB/s |

| Shading Units | 768 | 1664 |

| TMUs | 64 | 104 |

| ROPs | 16 | 56 |

| Pixel Rate | 14.85 GPixel/s | 65.97 GPixel/s |

| Texture Rate | 59.39 GTexel/s | 122.5 GTexel/s |

| FP32 Performance | 1,425.4 GFLOPS | 3.920 TFLOPS |

| TDP | 110 W | 148 W |

| Slot Width | Single-slot | Dual-slot |

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

| Display Outputs | 2x DVI, 1x mini-HDMI 1.4a | 1x DVI, 1x HDMI 2.0, 3x DisplayPort 1.2 |

| DirectX Support | 12 (11_0) | 12 (12_1) |

| Vulkan Support | 1.2.175 | 1.4 |

| Length | 145 mm, 5.7 inches | 267 mm, 10.5 inches |

| Height | Not recorded | 111 mm, 4.4 inches |

| Width | Not recorded | 40 mm, 1.6 inches |

| Release Date | 2012-10-08 | 2014-09-18 |

| Launch MSRP | 149 USD | 329 USD |

| Predecessor | GeForce 500 | GeForce 700 |

| Successor | GeForce 700 | GeForce 10 |

Fields that match include manufacturer (NVIDIA), process node (28 nm), foundry (TSMC), memory type (GDDR5), bus interface (PCIe 3.0 x16), OpenGL version (4.6), suggested PSU (300 W), and production status (end-of-life). Neither card has recorded base clocks for the GTX 650 Ti, game clocks, RT cores, tensor cores, or FP16 performance. The GTX 970 has additional recorded dimensions for height and width that the GTX 650 Ti lacks. The GTX 970 also has a Geekbench Metal score and a 3DMark Steel Nomad DX12 score, while the GTX 650 Ti has no recorded scores for those tests.