AMD Radeon RX 9060 vs NVIDIA GeForce GTX 660 Ti Comparison

The AMD Radeon RX 9060 and NVIDIA GeForce GTX 660 Ti represent two vastly different eras of GPU design, and the benchmark data reflects a generational chasm. The RX 9060, built on RDNA 4.0, utterly dominates the legacy Kepler-based GTX 660 Ti in every measurable test. In the two shared benchmarks, the RX 9060 posts an average score of 16,014 across all tests, landing in the 59th percentile of all GPUs, while the GTX 660 Ti's average of 15,063 places it in the 57th percentile. The most striking result is in Geekbench OpenCL, where the RX 9060 scores 88,183 against the GTX 660 Ti's 15,113, a delta of 483.5%. This is not a close contest; it is a complete overhaul of performance expectations.

Head-to-Head Benchmarks

The only two benchmarks both cards share are Geekbench OpenCL and Geekbench Vulkan, and the RX 9060 wins both outright. In OpenCL, the RX 9060's 88,183 score is roughly 5.8 times higher than the GTX 660 Ti's 15,113. That 483.5% delta highlights how far compute performance has come—the GTX 660 Ti's Kepler architecture simply cannot keep up with the parallel throughput of RDNA 4.0. For Vulkan, the RX 9060 scores 39,476 versus 15,012, a 163% advantage. While the gap narrows compared to OpenCL, it remains a landslide victory, showing the RX 9060 handles modern graphics APIs far more efficiently.

Looking at the RX 9060's broader benchmark suite, it achieves 3,322 in 3DMark Steel Nomad DX12, which is a solid result for a card with 1,792 shading units. Its Passmark G3D score of 17,631 and GPU compute score of 9,919 further cement its position. The GTX 660 Ti lacks these results, but its lower OpenCL and Vulkan scores suggest it would trail severely in DX12 and compute workloads. The RX 9060's nearest rival, the RTX 3060 Ti, averages 16,129, just 0.7% higher, meaning the RX 9060 sits right at the edge of modern mid-range performance. Meanwhile, the GTX 660 Ti's closest competitor, the RX 7600, averages 15,171, only 0.7% higher, placing the old card near modern entry-level territory despite its age.

Architecture Differences

The fundamental split here is architectural. The RX 9060 uses the Navi 44 chip on RDNA 4.0, fabricated on a 4 nm process at TSMC. It packs 29,700 million transistors into a 199 mm² die, yielding a transistor density of 149.2 million per square millimeter. In contrast, the GTX 660 Ti uses the GK104 chip on the Kepler architecture, built on a 28 nm process. It has just 3,540 million transistors spread across a much larger 294 mm² die, resulting in 12.0 million transistors per square millimeter. This density difference is the core reason for the performance gap—the RX 9060 crams over 12 times more transistors into a physically smaller package.

Memory configurations diverge sharply. The RX 9060 features 8 GB of GDDR6 on a 128-bit bus, delivering 288.0 GB/s of bandwidth. The GTX 660 Ti offers only 2 GB of GDDR5 on a 192-bit bus, with 144.2 GB/s bandwidth. While the GTX 660 Ti has a wider bus, the GDDR5's slower speed and lower capacity cripple it in modern texture-heavy scenes. The RX 9060 also brings 28 dedicated ray tracing cores, a feature entirely absent from the GTX 660 Ti, which has no RT cores or tensor cores at all. The RX 9060's FP32 throughput of 21.43 TFLOPS dwarfs the GTX 660 Ti's 2.634 TFLOPS, an 8-fold difference in raw compute.

FAQ

Q: Which card has better raw compute performance in OpenCL?

A: The AMD Radeon RX 9060 dominates, scoring 88,183 in Geekbench OpenCL compared to the GTX 660 Ti's 15,113, a 483.5% advantage. This reflects the RX 9060's modern 4 nm process and 21.43 TFLOPS FP32 throughput versus the GTX 660 Ti's 2.634 TFLOPS.

Q: Is the GTX 660 Ti still competitive in Vulkan games?

A: No. The RX 9060 scores 39,476 in Geekbench Vulkan versus the GTX 660 Ti's 15,012, a 163% gap. The GTX 660 Ti supports Vulkan 1.2.175, but its Kepler architecture lacks the hardware features needed for modern Vulkan workloads.

Q: How do their memory capacities affect performance?

A: The RX 9060's 8 GB GDDR6 offers 288.0 GB/s bandwidth, while the GTX 660 Ti's 2 GB GDDR5 provides only 144.2 GB/s. The RX 9060's higher capacity and bandwidth allow it to handle larger textures and higher resolutions without stuttering.

Q: What are the power requirements for each card?

A: The RX 9060 has a TDP of 132 W and requires a 300 W PSU with a single 8-pin connector. The GTX 660 Ti has a 150 W TDP and needs a 450 W PSU with two 6-pin connectors. Despite being older, the GTX 660 Ti consumes more power per unit of performance.

Q: Do both cards support ray tracing?

A: No. The RX 9060 includes 28 RT cores for hardware-accelerated ray tracing. The GTX 660 Ti has no RT cores or tensor cores, so it cannot handle ray-traced effects without severe performance penalties.

Q: Which card is better for legacy DirectX titles?

A: The RX 9060 supports DirectX 12 Ultimate (12_2) and scores 280 in Passmark DirectX 9, 182 in DX11, and 104 in DX10. The GTX 660 Ti only supports DirectX 12 (11_0) and lacks these Passmark scores, but its older architecture suggests it would fare worse in all API levels.

Specification Differences

The two cards differ in nearly every specification category. The RX 9060 uses a 4 nm process versus the GTX 660 Ti's 28 nm, and its transistor count of 29,700 million is over 8 times the GTX 660 Ti's 3,540 million. Die size favors the RX 9060 as well: 199 mm² versus 294 mm². Clock speeds are drastically different—the RX 9060 boosts to 2990 MHz with a base of 1700 MHz, while the GTX 660 Ti boosts to 980 MHz from a 915 MHz base. Memory sees the RX 9060 with 8 GB GDDR6 at 18 Gbps effective, whereas the GTX 660 Ti has 2 GB GDDR5 at 6 Gbps effective. The RX 9060's 128-bit bus is narrower than the GTX 660 Ti's 192-bit, but its bandwidth of 288.0 GB/s still doubles the GTX 660 Ti's 144.2 GB/s.

Shading units differ: 1,792 on the RX 9060 versus 1,344 on the GTX 660 Ti, though both have 112 TMUs. ROPs heavily favor the RX 9060 at 64 versus 24. Pixel rates are 191.4 GPixel/s for the RX 9060 against 27.44 GPixel/s for the GTX 660 Ti, and texture rates are 334.9 GTexel/s versus 109.8 GTexel/s. The RX 9060 supports PCIe 5.0 x16, while the GTX 660 Ti uses PCIe 3.0 x16. Display outputs differ, with the RX 9060 offering HDMI 2.1b and DisplayPort 2.1a, versus the GTX 660 Ti's DVI, HDMI 1.4a, and DisplayPort 1.2. The GTX 660 Ti has a launch MSRP of 299 USD, but the RX 9060 has no listed launch MSRP.

Where Each One Wins

The AMD Radeon RX 9060 wins in every category where data exists. It excels in modern workloads: ray tracing, high-bandwidth textures, and compute-heavy tasks like OpenCL and Vulkan. Its 8 GB VRAM and 288.0 GB/s bandwidth make it suitable for 1440p gaming and content creation. The 28 RT cores enable features the GTX 660 Ti cannot touch. The RX 9060's Passmark G3D score of 17,631 places it near the RTX 3060 Ti, which averages 16,129, so it competes with newer mid-range cards. Its 59th percentile ranking shows it sits above most GPUs in the database.

The NVIDIA GeForce GTX 660 Ti has no wins in the head-to-head data. Its only advantage is theoretical: a wider 192-bit memory bus and lower TDP of 150 W, though the RX 9060's 132 W TDP makes it more efficient overall. The GTX 660 Ti's 2 GB VRAM is insufficient for modern games, and its lack of RT cores and low FP32 throughput (2.634 TFLOPS) relegate it to legacy titles or basic desktop use. Its 57th percentile ranking is barely below the RX 9060, but that percentage is skewed by a much smaller benchmark sample size—the GTX 660 Ti only has two Geekbench results, while the RX 9060 has ten.

The Verdict

The data is unambiguous: the AMD Radeon RX 9060 is the superior card in every measured metric. With 483.5% higher OpenCL performance and 163% higher Vulkan performance, it is not just faster—it is in a different performance class. The RX 9060's 4 nm process, 29,700 million transistors, and 21.43 TFLOPS FP32 output represent a decade of architectural progress over the GTX 660 Ti's 28 nm Kepler design. For anyone building a system today, the RX 9060 is the only rational choice, offering modern API support, ray tracing, and double the memory bandwidth.

The GTX 660 Ti remains a historical artifact. Its 2 GB GDDR5 and 144.2 GB/s bandwidth cannot handle modern game assets, and its Vulkan 1.2.175 support is outdated. While its 150 W TDP is modest, the RX 9060's 132 W TDP is lower, making the newer card more efficient despite being dramatically faster. The GTX 660 Ti's 57th percentile ranking, achieved with just two benchmarks, flatters it—real-world performance in current titles would be far worse. If you are upgrading from a GTX 660 Ti, the RX 9060 delivers a 483.5% improvement in compute tasks alone, making it a transformative upgrade. For those starting fresh, the RX 9060's 59th percentile standing and proximity to the RTX 3060 Ti (0.7% delta) make it a practical mid-range pick, while the GTX 660 Ti should only be considered for retro builds or basic office work. The verdict is simple: buy the RX 9060; the GTX 660 Ti belongs in a museum.