AMD Radeon 660M vs AMD Radeon RX 9060 Comparison

The AMD Radeon RX 9060 and the AMD Radeon 660M are not in the same performance class, and the benchmark data confirms this decisively. The RX 9060 wins both recorded head-to-head tests by massive margins, with a 584.9% advantage in OpenCL and a 167.7% advantage in Vulkan. The 660M is an integrated graphics processor designed for mobile efficiency, while the RX 9060 is a discrete, actively cooled desktop part. The recorded data shows no scenario where the 660M comes out ahead; the question is not whether the RX 9060 wins, but by how much, and which specific workloads favor each architecture's design goals.

Head-to-Head Benchmarks

The direct comparison tests show an enormous gap between the two GPUs. In Geekbench OpenCL, the RX 9060 scores 88,183 points against 12,876 for the 660M, a delta of 584.9%. This is not a marginal improvement; it represents a completely different tier of compute throughput. The RX 9060's raw FP32 output of 21.43 TFLOPS dwarfs the 660M's 1,459.2 GFLOPS, and that 14.7x difference in theoretical peak translates directly into the measured OpenCL result.

In Geekbench Vulkan, the RX 9060 scores 39,476 against 14,748, a 167.7% lead. While still a commanding victory, the smaller percentage delta compared to OpenCL suggests the 660M's RDNA 2.0 architecture handles Vulkan's lower-level API overhead relatively better than it handles OpenCL's more general compute workloads. Still, the absolute numbers leave no ambiguity: the RX 9060 delivers more than 2.6 times the Vulkan performance of the 660M.

Looking at the broader benchmark picture, the RX 9060's average benchmark score is 16,014, while the 660M averages 13,812. That 16% overall advantage narrows the story considerably from the head-to-head tests, but the reason is clear: the 660M's recorded benchmarks are limited to the two Geekbench tests, whereas the RX 9060's average includes 10 different tests, some of which (like Passmark DirectX 9 at 280 and DirectX 12 at 44) are legacy or specialized workloads that drag down its aggregate. The 660M's percentile ranking of 55 versus the RX 9060's 59 reflects this compressed average, but the head-to-head data is the more reliable indicator of relative capability in modern workloads.

The nearest rivals for each GPU further illustrate the positioning. The RX 9060 sits within 2.1% of the AMD Radeon Pro W5500 (delta 2.1%), within 1% of the AMD Radeon RX 7700, and essentially tied with the NVIDIA GeForce RTX 3060 Ti (delta -0.7%). The 660M, by contrast, trades blows with the NVIDIA RTX A2000 Mobile (delta -0.1%) and the AMD Radeon RX 570X (delta -0.4%), and even shows a 0.5% edge over the AMD Radeon RX 7900 XT in this specific aggregate metric. These rival clusters confirm that the RX 9060 is competing with mid-range desktop discrete GPUs, while the 660M is competing with low-power mobile and older entry-level parts.

Architecture Differences

The architectural gap between these two GPUs spans a full generation and multiple design philosophies. The RX 9060 uses RDNA 4.0 architecture on a 4 nm TSMC process node, while the 660M uses RDNA 2.0 on a 6 nm TSMC node. The RX 9060's chip is the Navi 44 with 29,700 million transistors on a 199 mm² die, yielding a transistor density of 149.2M per mm². The 660M's Rembrandt chip contains 13,100 million transistors on a 208 mm² die, a density of just 63.0M per mm². Despite the 660M having a slightly larger physical die, the RX 9060 packs more than twice the transistors into a smaller area, a direct result of the newer process node.

The compute resources differ by roughly an order of magnitude. The RX 9060 has 1,792 shading units, 112 texture mapping units, and 64 ROPs, versus 384 shading units, 24 TMUs, and 16 ROPs for the 660M. Ray tracing cores follow the same pattern: 28 on the RX 9060 against 6 on the 660M. The pixel rate of the RX 9060 is 191.4 GPixel/s versus 30.40 GPixel/s for the 660M, and texture rate is 334.9 GTexel/s versus 45.60 GTexel/s. These are not incremental differences; they are structural ones.

Memory configuration is another fundamental divider. The RX 9060 has 8 GB of dedicated GDDR6 memory on a 128 bit bus with 288.0 GB/s of bandwidth and a memory clock of 2250 MHz (18 Gbps effective). The 660M uses System Shared memory, meaning its bandwidth is system dependent and its memory type and bus width are also system dependent. This means the 660M's performance is inherently tied to the host system's RAM speed and allocation, while the RX 9060 has its own dedicated, predictable memory subsystem.

Clock behavior also differs. The RX 9060 has a base clock of 1700 MHz, a game clock of 2400 MHz, and a boost clock of 2990 MHz. The 660M has a base clock of 1500 MHz and a boost clock of 1900 MHz, with no game clock specified. The RX 9060's higher clocks compound with its larger shader count to produce the massive FP32 advantage: 21.43 TFLOPS versus 1,459.2 GFLOPS. The FP16 numbers tell a different architectural story: the RX 9060 runs FP16 at a 1:1 ratio with FP32 (21.43 TFLOPS), while the 660M runs FP16 at a 2:1 ratio (2.918 TFLOPS). This means the RX 9060 does not gain a throughput advantage from FP16 workloads, whereas the 660M can double its throughput when using FP16 instructions.

Power and physical design further separate them. The RX 9060 has a TDP of 132 W, requires a 300 W suggested PSU, uses a dual-slot cooler with a single 8-pin power connector, and connects via PCIe 5.0 x16. The 660M is an IGP with a 40 W TDP, no power connectors, no slot width (it is integrated into the processor), and uses PCIe 4.0 x8. The RX 9060 outputs to 1x HDMI 2.1b and 2x DisplayPort 2.1a, while the 660M's display outputs are portable device dependent. The RX 9060 is marked as Active production, released 2025-08-04, while the 660M is End-of-life, released 2022-01-03, with a successor listed as Navi III IGP.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon RX 9060 has an average benchmark score of 16,014, while the AMD Radeon 660M averages 13,812. The RX 9060 also ranks in the 59th percentile of all GPUs, versus the 55th percentile for the 660M.

Q: How large is the performance gap in the head-to-head tests?

A: In Geekbench OpenCL, the RX 9060 leads by 584.9% (88,183 versus 12,876). In Geekbench Vulkan, the RX 9060 leads by 167.7% (39,476 versus 14,748). The RX 9060 wins both recorded head-to-head tests.

Q: What memory does each GPU use?

A: The RX 9060 uses 8 GB of dedicated GDDR6 memory on a 128 bit bus with 288.0 GB/s bandwidth. The 660M uses System Shared memory, with system dependent bandwidth and no fixed memory size or bus width.

Q: What are the process nodes for these chips?

A: The RX 9060 uses a 4 nm TSMC process node with 29,700 million transistors on a 199 mm² die. The 660M uses a 6 nm TSMC node with 13,100 million transistors on a 208 mm² die. The RX 9060 achieves 149.2M transistors per mm² compared to 63.0M per mm² for the 660M.

Q: What is the power consumption difference?

A: The RX 9060 has a TDP of 132 W with a dual-slot cooler and a 1x 8-pin power connector, requiring a 300 W suggested PSU. The 660M has a TDP of 40 W, uses no power connectors, and is an IGP with no slot width.

Q: How do their nearest rivals compare?

A: The RX 9060's nearest rival is the NVIDIA GeForce RTX 3060 Ti, with a delta of -0.7%, and it also sits within 1% of the AMD Radeon RX 7700. The 660M's nearest rival is the NVIDIA RTX A2000 Mobile with a delta of -0.1%, and it is 0.5% ahead of the AMD Radeon RX 7900 XT in average score.

The Verdict

The data supports only one conclusion for performance: the AMD Radeon RX 9060 is the superior GPU in every measured category. Its average benchmark score is 16% higher, its head-to-head wins are by 584.9% and 167.7%, and its hardware specifications (1792 shading units versus 384, 8 GB GDDR6 versus system shared memory, 21.43 TFLOPS versus 1,459.2 GFLOPS) show a product designed for a completely different workload tier. The 660M's only advantages are its 40 W TDP versus 132 W, its IGP form factor that requires no separate power connectors, and its smaller physical footprint as an integrated solution. For any user who needs discrete-level gaming or compute performance, the RX 9060 is the only choice between these two. The 660M makes sense only for ultra-portable or low-power systems where the 132 W TDP and dual-slot cooler of the RX 9060 are physically impossible to accommodate. The production statuses reinforce this: the RX 9060 is Active, while the 660M is End-of-life. A builder choosing between these should treat the 660M as a legacy integrated option and the RX 9060 as a current discrete solution.

Specification Differences

| Specification | AMD Radeon RX 9060 | AMD Radeon 660M |

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

| Architecture | RDNA 4.0 | RDNA 2.0 |

| Process Node | 4 nm | 6 nm |

| Transistors | 29,700 million | 13,100 million |

| Die Size | 199 mm² | 208 mm² |

| Transistor Density | 149.2M / mm² | 63.0M / mm² |

| Base Clock | 1700 MHz | 1500 MHz |

| Boost Clock | 2990 MHz | 1900 MHz |

| Game Clock | 2400 MHz | None |

| Memory Size | 8 GB | System Shared |

| Memory Type | GDDR6 | System Shared |

| Memory Bus | 128 bit | System Shared |

| Memory Bandwidth | 288.0 GB/s | System Dependent |

| Memory Clock | 2250 MHz (18 Gbps effective) | System Shared |

| Shading Units | 1792 | 384 |

| TMUs | 112 | 24 |

| ROPs | 64 | 16 |

| RT Cores | 28 | 6 |

| Pixel Rate | 191.4 GPixel/s | 30.40 GPixel/s |

| Texture Rate | 334.9 GTexel/s | 45.60 GTexel/s |

| FP32 | 21.43 TFLOPS | 1,459.2 GFLOPS |

| FP16 | 21.43 TFLOPS (1:1) | 2.918 TFLOPS (2:1) |

| TDP | 132 W | 40 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 1x 8-pin | None |

| Suggested PSU | 300 W | None |

| Bus Interface | PCIe 5.0 x16 | PCIe 4.0 x8 |

| Display Outputs | 1x HDMI 2.1b, 2x DisplayPort 2.1a | Portable Device Dependent |

| Production Status | Active | End-of-life |

| Release Date | 2025-08-04 | 2022-01-03 |

| Predecessor | Navi III | Vega II IGP |

| Successor | None | Navi III IGP |

Where Each One Wins

The RX 9060 wins every performance-oriented category. It dominates in raw compute throughput (21.43 TFLOPS FP32 versus 1,459.2 GFLOPS), memory bandwidth (288.0 GB/s dedicated versus system dependent shared memory), pixel fill rate (191.4 GPixel/s versus 30.40 GPixel/s), and texture fill rate (334.9 GTexel/s versus 45.60 GTexel/s). Its 8 GB of dedicated GDDR6 memory ensures consistent performance independent of the host system's RAM configuration. The ray tracing core count (28 versus 6) gives it a substantial advantage in RT workloads. For gaming at 1080p or higher, for content creation with GPU acceleration, or for any compute-heavy task, the RX 9060 is the clear winner.

The 660M wins only in efficiency and integration categories. Its 40 W TDP is less than one-third of the RX 9060's 132 W, making it suitable for thin-and-light laptops and low-power mini PCs where the RX 9060's dual-slot cooler and 1x 8-pin power connector would be impossible to install. It requires no separate power connectors and no suggested PSU, and its PCIe 4.0 x8 interface is sufficient for an integrated part. Its 6 nm node and 208 mm² die are larger than the RX 9060's 4 nm and 199 mm², but this is because it shares the die with the host processor. The 660M also benefits from FP16 2:1 ratio execution, which gives it a relative efficiency advantage in workloads that can use FP16 instructions, though the RX 9060's absolute FP16 throughput is still far higher at 21.43 TFLOPS versus 2.918 TFLOPS. For a user building a new system, the RX 9060 is the only sensible choice for performance; the 660M is a legacy integrated option for systems where physical space and power draw trump all other considerations.