AMD Radeon RX 470 vs AMD Radeon RX 6700 Comparison

AMD Radeon RX 6700 and AMD Radeon RX 470 are both end-of-life AMD graphics cards, but they belong to different architectural eras and performance tiers. The RX 6700 is a RDNA 2.0 part built on a 7 nm process, while the RX 470 is a GCN 4.0 part on a 14 nm process. Benchmark data shows the RX 6700 delivers a dominant performance advantage across every shared test, with the RX 470 holding a slight edge only in its own percentile ranking relative to all GPUs (74th vs. 75th percentile, a marginal difference of one point). The following analysis breaks down the head-to-head results, architectural divergence, and practical implications for each card.

Head-to-Head Benchmarks

The RX 6700 wins all four benchmarks where both cards have recorded scores, and the margins are substantial. In the 3DMark Steel Nomad DX12 test, the RX 6700 scores 2014 against the RX 470’s 842, a delta of 139.2% in favor of the newer card. This is the smallest relative victory in the set, yet it still represents more than double the raw score. The largest gap appears in Geekbench Metal, where the RX 6700 posts 122223 versus 41690, a 193.2% delta. That means the RX 6700 is nearly three times faster in Apple’s Metal API, a significant advantage for any compute or rendering workload using that interface.

Geekbench OpenCL shows a similar story: RX 6700 scores 97841, RX 470 scores 33568, for a 191.5% delta. The Vulkan result is comparatively closer but still lopsided: 83974 for the RX 6700 versus 39884 for the RX 470, a 110.5% delta. Across all four tests, the RX 6700’s smallest win is just over double the RX 470’s score (110.5% delta), and its largest is nearly triple (193.2% delta). The data shows no benchmark where the RX 470 is competitive; the RX 6700’s average benchmark score of 30433 further confirms this, sitting 1.6% above the NVIDIA GeForce RTX 3070 Ti (29945) and 1.1% above the AMD Radeon RX 6800 (30095) in its nearest rival group.

The RX 470’s average score of 28996 places it among a different competitive set, within 0.4% of the AMD Radeon RX 6800M (28874) and 0.7% below the AMD Radeon RX Vega M GH (29197). While the RX 470’s percentile rank (74th) is nearly identical to the RX 6700’s (75th), that ranking reflects its position against all GPUs, not against the RX 6700 directly. In direct comparison, the RX 6700 leads by 139% to 193% depending on the API, which dwarfs the narrow percentile difference. The head-to-head record is 4 wins for the RX 6700 and 0 for the RX 470, with no test where the older card even approaches parity.

Architecture Differences

The architectural gap between these two cards is generational. The RX 6700 uses the Navi 22 chip with RDNA 2.0 architecture, fabricated by TSMC on a 7 nm process with 17,200 million transistors on a 335 mm² die. The RX 470 uses the Ellesmere chip with GCN 4.0 architecture, fabricated by GlobalFoundries on a 14 nm process with 5,700 million transistors on a 232 mm² die. Transistor density tells the story: the RX 6700 packs 51.3 million transistors per square millimeter, versus 24.6 million for the RX 470 — more than double the density. This allows the RX 6700 to fit three times as many transistors (17,200 million vs. 5,700 million) into a die that is only 44% larger by area.

Shader and compute resources also diverge sharply. The RX 6700 has 2304 shading units, 144 texture mapping units (TMUs), and 64 render output units (ROPs). The RX 470 has 2048 shading units, 128 TMUs, and 32 ROPs. The RX 6700’s ROP count is exactly double, which contributes to its pixel rate of 156.8 GPixel/s versus 38.59 GPixel/s for the RX 470 — a 4x advantage. Texture rate is 352.8 GTexel/s versus 154.4 GTexel/s, a 2.3x advantage. Floating-point performance is 11.29 TFLOPS (FP32) versus 4.940 TFLOPS, a 2.3x gap. The RX 6700 also introduces 36 ray tracing cores, a feature entirely absent from the RX 470. The RX 6700’s FP16 rate is 22.58 TFLOPS (2:1 ratio), while the RX 470’s FP16 is 4.940 TFLOPS (1:1 ratio), meaning the newer card has a dedicated half-precision path.

Memory subsystems are fundamentally different. The RX 6700 uses 10 GB of GDDR6 on a 160-bit bus, delivering 320.0 GB/s of bandwidth. The RX 470 uses 4 GB of GDDR5 on a 256-bit bus, delivering 211.2 GB/s. Although the RX 470 has a wider bus, the RX 6700’s faster memory (16 Gbps effective vs. 6.6 Gbps effective) yields 51.5% more bandwidth. The RX 6700 also supports PCIe 4.0 x16, while the RX 470 is limited to PCIe 3.0 x16. API support differs: the RX 6700 supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the RX 470 supports DirectX 12 (12_0) and Vulkan 1.3. Both support OpenGL 4.6. The RX 6700’s power draw is 175 W with a 1x 8-pin connector and a 450 W suggested PSU; the RX 470 draws 120 W with a 1x 6-pin connector and a 300 W suggested PSU.

Where Each One Wins

The RX 6700 wins in every measured scenario, but the magnitude of the win varies by API and workload type. In Geekbench Metal and OpenCL, the deltas exceed 190%, indicating the RX 6700 is exceptionally strong in compute-heavy tasks that leverage these APIs — particularly relevant for content creation, GPU compute, and cross-platform workloads. The Vulkan delta of 110.5% is the smallest, suggesting the RX 470’s GCN architecture is relatively less disadvantaged in Vulkan than in other APIs, though still far behind. The 3DMark Steel Nomad DX12 test shows a 139.2% gap, placing the RX 6700 in a position where it can handle modern DX12 gaming at higher settings and resolutions, whereas the RX 470 would be limited to lower settings or older titles.

The RX 470’s only "win" is its power efficiency in absolute terms: it draws 120 W versus 175 W, and requires a 300 W PSU versus 450 W. In performance-per-watt, however, the RX 6700’s 11.29 TFLOPS at 175 W yields 64.5 GFLOPS per watt, versus the RX 470’s 41.2 GFLOPS per watt (4.940 TFLOPS at 120 W). The RX 470 also has a smaller physical footprint: 240 mm length, 95 mm height, 35 mm width versus the RX 6700’s 267 mm length, 110 mm height, 40 mm width. For a compact system with a low-wattage PSU, the RX 470 remains a viable legacy option; for any modern workload, the RX 6700 is categorically superior.

FAQ

Q: How much faster is the RX 6700 than the RX 470 in the 3DMark Steel Nomad DX12 test?

A: The RX 6700 scores 2014 versus the RX 470’s 842, a delta of 139.2% in favor of the RX 6700. This is the smallest relative gap among the four shared benchmarks.

Q: Which card has more memory bandwidth, and by how much?

A: The RX 6700 has 320.0 GB/s of bandwidth from 10 GB of GDDR6 on a 160-bit bus. The RX 470 has 211.2 GB/s from 4 GB of GDDR5 on a 256-bit bus. The RX 6700 leads by 51.5% despite the RX 470’s wider bus.

Q: Does the RX 470 support ray tracing?

A: No. The RX 470 has no ray tracing cores. The RX 6700 includes 36 ray tracing cores, a feature absent from the older GCN architecture.

Q: What is the difference in transistor density between the two cards?

A: The RX 6700 has a transistor density of 51.3 million per square millimeter, while the RX 470 has 24.6 million. The RX 6700’s density is more than double, enabled by the 7 nm process versus 14 nm.

Q: In which API does the RX 470 show the smallest performance gap relative to the RX 6700?

A: In Vulkan, the RX 470 scores 39884 against the RX 6700’s 83974, a 110.5% delta. This is the narrowest margin across the four shared benchmarks, though the RX 6700 still wins decisively.

Q: What are the power requirements for each card?

A: The RX 6700 has a TDP of 175 W, uses a 1x 8-pin power connector, and requires a 450 W suggested PSU. The RX 470 has a TDP of 120 W, uses a 1x 6-pin connector, and requires a 300 W suggested PSU.

The Verdict

The data is unambiguous: the AMD Radeon RX 6700 is the superior card for any performance-oriented task. It wins all four head-to-head benchmarks, with deltas ranging from 110.5% to 193.2%. Its average benchmark score of 30433 places it within 1.6% of the NVIDIA GeForce RTX 3070 Ti and 1.1% above the AMD Radeon RX 6800, indicating it competes with a much higher tier than the RX 470. The RX 6700’s architectural advantages — RDNA 2.0, 7 nm process, 17,200 million transistors, 36 ray tracing cores, 10 GB GDDR6, and 320.0 GB/s bandwidth — all contribute to this dominance. The RX 470, with GCN 4.0, 14 nm process, 5,700 million transistors, 4 GB GDDR5, and 211.2 GB/s bandwidth, is only competitive in terms of absolute power draw (120 W vs. 175 W) and physical dimensions (240 mm vs. 267 mm length).

For a user with a modern system and a 450 W PSU, the RX 6700 is the clear choice for gaming and compute workloads, especially in DirectX 12 Ultimate and Vulkan 1.4 contexts. For a legacy system constrained to a 300 W PSU or a compact chassis, the RX 470 remains a functional but dramatically slower option — its percentile rank (74th) is only one point below the RX 6700’s (75th), but that rank is misleading because the RX 6700’s nearest rivals are the RTX 3070 Ti and RX 6800, while the RX 470’s nearest rivals are the RX 6800M and Arc A370M. The RX 470’s launch MSRP was 179 USD, but no pricing data exists for the RX 6700. The verdict is simple: the RX 6700 is a generational leap, and the RX 470 is a legacy part that cannot match it in any benchmark.

Specification Differences

| Specification | AMD Radeon RX 6700 | AMD Radeon RX 470 |

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

| Architecture | RDNA 2.0 | GCN 4.0 |

| Process Node | 7 nm (TSMC) | 14 nm (GlobalFoundries) |

| Transistors | 17,200 million | 5,700 million |

| Die Size | 335 mm² | 232 mm² |

| Transistor Density | 51.3M / mm² | 24.6M / mm² |

| Base Clock | 1941 MHz | 926 MHz |

| Boost Clock | 2450 MHz | 1206 MHz |

| Game Clock | 2174 MHz | — |

| Memory Clock | 2000 MHz (16 Gbps effective) | 1650 MHz (6.6 Gbps effective) |

| Memory Size | 10 GB GDDR6 | 4 GB GDDR5 |

| Memory Bus Width | 160 bit | 256 bit |

| Memory Bandwidth | 320.0 GB/s | 211.2 GB/s |

| Shading Units | 2304 | 2048 |

| TMUs | 144 | 128 |

| ROPs | 64 | 32 |

| Ray Tracing Cores | 36 | — |

| Pixel Rate | 156.8 GPixel/s | 38.59 GPixel/s |

| Texture Rate | 352.8 GTexel/s | 154.4 GTexel/s |

| FP32 | 11.29 TFLOPS | 4.940 TFLOPS |

| FP16 | 22.58 TFLOPS (2:1) | 4.940 TFLOPS (1:1) |

| TDP | 175 W | 120 W |

| Power Connectors | 1x 8-pin | 1x 6-pin |

| Suggested PSU | 450 W | 300 W |

| Bus Interface | PCIe 4.0 x16 | PCIe 3.0 x16 |

| Display Outputs | 1x HDMI 2.1, 3x DisplayPort 1.4a | 1x HDMI 2.0b, 3x DisplayPort 1.4a |

| DirectX | 12 Ultimate (12_2) | 12 (12_0) |

| Vulkan | 1.4 | 1.3 |

| Dimensions (L×H×W) | 267×110×40 mm | 240×95×35 mm |

| Release Date | 2021-06-08 | 2016-08-03 |

| Launch MSRP | — | 179 USD |