AMD Radeon 860M vs AMD Radeon RX 6700M Comparison

# AMD Radeon 860M vs AMD Radeon RX 6700M

The AMD Radeon 860M and AMD Radeon RX 6700M represent two very different approaches to mobile graphics: the 860M is a 15 W integrated GPU built on TSMC's 4 nm process for Krackan Point, while the RX 6700M is a 135 W discrete part on the 7 nm Navi 22 chip with 10 GB of dedicated GDDR6 memory. Benchmark results show a decisive performance gap, with the RX 6700M winning both head-to-head tests by margins of 66.9% and 70.7%, yet the 860M holds a slightly higher overall percentile ranking (72 vs 71) due to differences in benchmark suites and scoring distributions.

FAQ

Q: Which GPU is faster in the shared head-to-head benchmarks?

A: The AMD Radeon RX 6700M wins both tests decisively. In Geekbench OpenCL it scores 77,666 versus the 860M's 22,759 (a -70.7% delta for the 860M), and in Geekbench Vulkan it scores 90,816 versus 30,043 (a -66.9% delta).

Q: How does the average benchmark score compare between the two?

A: The 860M has a higher average benchmark score of 26,401, while the RX 6700M averages 25,633. This is despite the RX 6700M winning the head-to-head tests, because the 860M's average is drawn from only two Geekbench results, whereas the RX 6700M's average includes eleven tests spanning 3DMark, PassMark, and Geekbench.

Q: What is the process node difference?

A: The 860M is fabricated on TSMC's 4 nm process, while the RX 6700M uses TSMC's 7 nm node. The RX 6700M has a known die size of 335 mm² and 17,200 million transistors with a density of 51.3M per mm²; the 860M's transistor count and die size are listed as unknown.

Q: What are the memory configurations?

A: The RX 6700M has 10 GB of GDDR6 on a 160-bit bus with 320.0 GB/s bandwidth. The 860M uses system-shared memory, with system-dependent bandwidth and no dedicated VRAM.

Q: Which GPU has a higher TDP?

A: The RX 6700M is rated at 135 W, which is nine times the 860M's 15 W TDP. The 860M is an integrated GPU with no power connectors, and the RX 6700M also lists no power connectors despite its higher power draw.

Q: What are the production statuses?

A: The 860M is listed as Active with a release date of 2025-02-28, while the RX 6700M is End-of-life and was released on 2021-05-30. The 860M's predecessor is Navi II IGP; the RX 6700M's predecessor is Polaris Mobile.

Architecture Differences

The two GPUs belong to different RDNA generations and are built for fundamentally different roles. The 860M uses RDNA 3.5 architecture on TSMC's 4 nm process, part of the Navi III IGP generation for Strix Point Mobile. The RX 6700M uses RDNA 2.0 on TSMC's 7 nm node, belonging to the Navi Mobile (RX 6000M) generation with the Navi 22 chip.

Shading resources differ substantially. The 860M has 512 shading units, 32 texture mapping units, and 16 raster operation pipelines. The RX 6700M has 2,304 shading units, 144 TMUs, and 64 ROPs — roughly 4.5 times the shader count, 4.5 times the TMUs, and 4 times the ROPs. Ray tracing cores also differ: the 860M has 8, while the RX 6700M has 36.

Clock behavior reflects the design philosophy. The 860M has a 600 MHz base clock that boosts to 3000 MHz, while the RX 6700M runs at a 1489 MHz base and 2400 MHz boost with a 2300 MHz game clock. The RX 6700M's memory clock is 2000 MHz with 16 Gbps effective, whereas the 860M's memory clock is system-shared.

Compute throughput is dramatically different. The 860M delivers 3.072 TFLOPS FP32 and 3.072 TFLOPS FP16 (1:1 ratio). The RX 6700M delivers 11.06 TFLOPS FP32 and 22.12 TFLOPS FP16 (2:1 ratio). Pixel and texture rates follow the same pattern: the 860M achieves 48.00 GPixel/s and 96.00 GTexel/s, while the RX 6700M reaches 153.6 GPixel/s and 345.6 GTexel/s.

Both GPUs support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The bus interface differs: the 860M uses PCIe 4.0 x8, while the RX 6700M uses PCIe 4.0 x16. Both are listed with IGP slot width and no power connectors, with display outputs described as portable device dependent.

Head-to-Head Benchmarks

The shared test suite contains only two benchmarks, and the RX 6700M dominates both. In Geekbench OpenCL, the RX 6700M scores 77,666 against the 860M's 22,759. The delta of -70.7% means the 860M achieves less than a third of the RX 6700M's score. In Geekbench Vulkan, the RX 6700M scores 90,816 versus 30,043, a -66.9% delta for the 860M.

These are large, consistent margins. The Vulkan gap is slightly narrower than OpenCL, but both remain in the same ballpark: the 860M trails by roughly two-thirds to three-quarters of the RX 6700M's output. The RX 6700M's advantage is not marginal or workload-specific; it is a structural lead across both APIs tested.

The 860M's average benchmark score of 26,401 is higher than the RX 6700M's 25,633, which is counterintuitive given the head-to-head results. This stems from the benchmark sets: the 860M's average comes only from Geekbench OpenCL (22,759) and Vulkan (30,043), while the RX 6700M's average includes PassMark DirectX 9 (155), DirectX 10 (92), DirectX 11 (129), DirectX 12 (65), G2D (555), G3D (13,536), and GPU Compute (5,191), plus Geekbench Metal (91,911), OpenCL (77,666), and Vulkan (90,816), and 3DMark Steel Nomad DX12 (1,845). The heavy weight of the 91,911 Metal score and the 90,816 Vulkan score pulls the RX 6700M's average up, but the low PassMark sub-scores drag it down. The 860M's two-score average benefits from having no low-end tests included.

Nearest rival data puts both in the same performance tier. The 860M's closest rivals are the NVIDIA GeForce MX550 (average 26,421, -0.1% delta), NVIDIA GeForce RTX 5060 (26,331, +0.3%), AMD Radeon RX 5700 XT 50th Anniversary (26,553, -0.6%), and NVIDIA RTX A4000 (26,683, -1.1%). The RX 6700M's rivals are the AMD Radeon Pro W5700 (25,726, -0.4%), NVIDIA GeForce RTX 3080 Ti Mobile (25,740, -0.4%), AMD FirePro D700 (25,842, -0.8%), and AMD FirePro W7100 (25,856, -0.9%). Both GPUs sit within 1.1% of their nearest competitors, indicating that average scores alone do not separate them from the field.

Specification Differences

| Specification | AMD Radeon 860M | AMD Radeon RX 6700M |

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

| Architecture | RDNA 3.5 | RDNA 2.0 |

| Process node | 4 nm | 7 nm |

| Chip | Krackan Point | Navi 22 |

| Generation | Navi III IGP (Strix Point Mobile) | Navi Mobile (RX 6000M) |

| Transistors | unknown | 17,200 million |

| Die size | unknown | 335 mm² |

| Transistor density | null | 51.3M / mm² |

| Base clock | 600 MHz | 1489 MHz |

| Boost clock | 3000 MHz | 2400 MHz |

| Game clock | null | 2300 MHz |

| Memory size | System Shared | 10 GB |

| Memory type | System Shared | GDDR6 |

| Memory bus | System Shared | 160 bit |

| Memory bandwidth | System Dependent | 320.0 GB/s |

| Shading units | 512 | 2304 |

| TMUs | 32 | 144 |

| ROPs | 16 | 64 |

| RT cores | 8 | 36 |

| Pixel rate | 48.00 GPixel/s | 153.6 GPixel/s |

| Texture rate | 96.00 GTexel/s | 345.6 GTexel/s |

| FP32 | 3.072 TFLOPS | 11.06 TFLOPS |

| FP16 | 3.072 TFLOPS (1:1) | 22.12 TFLOPS (2:1) |

| TDP | 15 W | 135 W |

| Bus interface | PCIe 4.0 x8 | PCIe 4.0 x16 |

| Production status | Active | End-of-life |

| Release date | 2025-02-28 | 2021-05-30 |

| Predecessor | Navi II IGP | Polaris Mobile |

Where Each One Wins

The RX 6700M wins every head-to-head benchmark against the 860M. In Geekbench OpenCL it is 70.7% ahead, and in Vulkan it is 66.9% ahead. These are the only direct comparisons available, and they show no scenario where the 860M takes the lead. The RX 6700M's 3.6x advantage in FP32 throughput (11.06 vs 3.072 TFLOPS), 3.33x higher pixel rate, and 3.6x higher texture rate all point to consistent superiority in raw compute-bound tasks. Its 10 GB GDDR6 with 320.0 GB/s bandwidth versus system-shared memory also gives it a decisive edge in memory-heavy workloads such as high-resolution textures or large datasets.

The 860M's wins are not in performance but in efficiency and integration. Its 15 W TDP is one-ninth of the RX 6700M's 135 W, making it suitable for thin-and-light portable devices where the RX 6700M would be impractical. The 860M is also on a newer 4 nm process versus 7 nm, and it remains in Active production while the RX 6700M is End-of-life. The 860M's higher average benchmark score (26,401 vs 25,633) and higher percentile ranking (72 vs 71) are notable, but these reflect the different benchmark suites rather than a real performance advantage — the RX 6700M's average is depressed by PassMark sub-scores that the 860M never takes.

For use-case planning: the RX 6700M is the choice for any workload that demands sustained frame rates, high resolutions, or compute throughput — gaming, 3D rendering, or GPU-accelerated processing where the 11.06 TFLOPS and 320 GB/s bandwidth can be utilized. The 860M is the choice for ultra-portable systems where power draw and thermal output are the primary constraints, and where its 3.072 TFLOPS and system-shared memory are sufficient for lighter tasks. The data does not support any scenario where the 860M outperforms the RX 6700M in a direct benchmark comparison; the RX 6700M's wins in both head-to-head tests are decisive and uncontested.