AMD Radeon 680M vs AMD Radeon Pro 460 Comparison

Where Each One Wins

The benchmark data splits these two integrated GPUs cleanly, but not evenly. The AMD Radeon 680M wins every head-to-head comparison recorded in the database. In the two shared tests, it takes both victories decisively. The AMD Radeon Pro 460 does not win a single direct comparison, yet its average benchmark score tells a different story.

Look at the aggregate numbers. The Radeon Pro 460 has an average benchmark score of 17,509 across all recorded tests. The Radeon 680M sits at 15,270. That is a 14.7% gap in favor of the older part, despite the 680M winning both direct matchups. This happens because the two GPUs were tested under different workloads. The Pro 460 was pushed through Geekbench Metal, OpenCL, and Vulkan tests. The 680M has results for 3DMark Steel Nomad DX12, Geekbench OpenCL, and Geekbench Vulkan. The 3DMark result is far lower than anything else in the pack, dragging the 680M average down.

For compute-heavy tasks using OpenCL or Vulkan, the 680M is the clear winner. Its OpenCL score of 23,468 beats the Pro 460's 15,284 by 34.9%. In Vulkan, the 680M scores 21,965 against 16,816, a 23.4% advantage. If your workload uses those APIs, pick the 680M without hesitation.

The Pro 460 counters with a Metal score of 20,426. The 680M has no Metal result in the database. For Apple-centric workflows that favor Metal, the Pro 460 has a measurable advantage, though the data only shows its absolute score, not a direct comparison. The Pro 460 also sits at the 61st percentile of all GPUs, while the 680M is at the 57th. That percentile ranking reflects the broader test suite, where the Pro 460's consistent scores across three APIs give it a higher standing.

Architecture Differences

These two GPUs come from different eras and different design philosophies. The Radeon Pro 460 uses the Baffin chip built on GCN 4.0 architecture. It is manufactured by GlobalFoundries on a 14 nm process. The chip packs 3,000 million transistors into a 123 mm² die, yielding a transistor density of 24.4 million per square millimeter. This is a discrete-class GPU from 2016, though it ships in integrated form for Mac notebooks.

The Radeon 680M uses the Rembrandt+ chip based on RDNA 2.0 architecture. TSMC fabricates it on a 6 nm node. The chip contains 13,100 million transistors on a 208 mm² die, for a density of 63.0 million per square millimeter. That is more than 4 times the transistor count and nearly 2.6 times the density of the Pro 460. This is a modern APU integrated GPU from 2023.

The architectural differences run deeper than the node. GCN 4.0 uses a scalar and vector unit design that has aged poorly for modern workloads. RDNA 2.0 introduces a redesigned compute unit with better instruction efficiency. The 680M includes 12 ray tracing cores, a feature the Pro 460 lacks entirely. The 680M also supports DirectX 12 Ultimate (12_2), while the Pro 460 tops out at DirectX 12 (12_0). Vulkan support differs too: the 680M supports Vulkan 1.4, the Pro 460 only Vulkan 1.3.

Shader configuration diverges significantly. The Pro 460 has 1,024 shading units, 64 texture mapping units, and 16 ROPs. The 680M has fewer shading units at 768, fewer TMUs at 48, but double the ROPs at 32. The 680M compensates for lower shader count with much higher clocks, as the specification table shows.

Head-to-Head Benchmarks

The direct comparisons in the database cover two tests: Geekbench OpenCL and Geekbench Vulkan. The 680M wins both.

In Geekbench OpenCL, the 680M scores 23,468 against the Pro 460's 15,284. That is a 34.9% difference. This is the single largest gap between the two parts. The 680M's FP32 throughput of 3.379 TFLOPS explains this. The Pro 460 manages only 1.858 TFLOPS. The 680M delivers 81.8% more raw compute, and that translates directly into the OpenCL result.

In Geekbench Vulkan, the 680M scores 21,965 against 16,816. The 23.4% delta is smaller but still substantial. Vulkan benefits from the 680M's modern architecture and higher pixel rate. The 680M pushes 70.40 GPixel/s compared to the Pro 460's 14.51 GPixel/s. Texture rate also favors the 680M: 105.6 GTexel/s versus 58.05 GTexel/s. These throughput advantages show up in Vulkan workloads that stress rasterization and texture sampling.

The 680M also has a 3DMark Steel Nomad DX12 score of 378, but the Pro 460 has no comparable result. This test exercises DirectX 12 ray tracing and mesh shaders, features the Pro 460 cannot run at full capability given its older DirectX 12_0 support.

When you look at the broader rival landscape, the Pro 460's average score of 17,509 puts it within 0.5% of the AMD Radeon 780M (17,588) and 0.7% below the NVIDIA GeForce RTX 4060 (17,639). The 680M's average of 15,270 lands it within 0.5% of the NVIDIA GeForce RTX 3050 OEM (15,199) and 0.7% above the AMD Radeon RX 7600 (15,171). These rival comparisons are based on average scores across different test suites, so interpret them as general positioning rather than direct apples-to-apples matchups.

Specification Differences

The two GPUs differ in nearly every measurable specification. Here are the fields where they diverge:

• Chip: Baffin versus Rembrandt+
• Architecture: GCN 4.0 versus RDNA 2.0
• Generation: Radeon Pro Mac (400 Series) versus Navi II IGP (Rembrandt Mobile)
• Process node: 14 nm versus 6 nm
• Foundry: GlobalFoundries versus TSMC
• Transistors: 3,000 million versus 13,100 million
• Die size: 123 mm² versus 208 mm²
• Transistor density: 24.4M / mm² versus 63.0M / mm²
• Base clock: 850 MHz versus 2000 MHz
• Boost clock: 907 MHz versus 2200 MHz
• Memory type: GDDR5 versus System Shared
• Memory size: 4 GB versus System Shared
• Memory bus width: 128 bit versus System Shared
• Memory bandwidth: 81.28 GB/s versus System Dependent
• Shading units: 1024 versus 768
• TMUs: 64 versus 48
• ROPs: 16 versus 32
• Ray tracing cores: None versus 12
• Pixel rate: 14.51 GPixel/s versus 70.40 GPixel/s
• Texture rate: 58.05 GTexel/s versus 105.6 GTexel/s
• FP32: 1.858 TFLOPS versus 3.379 TFLOPS
• FP16: 1.858 TFLOPS (1:1) versus 6.758 TFLOPS (2:1)
• TDP: 35 W versus 50 W
• Bus interface: PCIe 3.0 x8 versus PCIe 4.0 x8
• DirectX support: 12 (12_0) versus 12 Ultimate (12_2)
• Vulkan support: 1.3 versus 1.4
• Release date: 2016-10-29 versus 2023-01-02
• Production status: End-of-life versus Active
• Predecessor: None versus Vega II IGP
• Successor: None versus Navi III IGP

The 680M has a clear spec advantage in clocks, memory bandwidth (though system dependent), pixel rate, texture rate, FP32, and FP16. The Pro 460 counters with dedicated 4 GB GDDR5 memory, which avoids the system memory bottleneck that can affect integrated GPUs. The 680M's FP16 throughput of 6.758 TFLOPS is particularly notable, running at a 2:1 ratio versus FP32, while the Pro 460 runs FP16 at a 1:1 ratio with only 1.858 TFLOPS.

FAQ

Q: Which GPU is faster in OpenCL workloads?

A: The AMD Radeon 680M scores 23,468 in Geekbench OpenCL, which is 34.9% higher than the Radeon Pro 460's 15,284.

Q: Does the Radeon Pro 460 support ray tracing?

A: No. The Pro 460 has no ray tracing cores. The Radeon 680M includes 12 ray tracing cores and supports DirectX 12 Ultimate (12_2).

Q: Which GPU has a higher average benchmark score?

A: The Radeon Pro 460 has an average score of 17,509 across its recorded tests, while the Radeon 680M averages 15,270. The 680M's lower average is influenced by its 3DMark Steel Nomad DX12 score of 378.

Q: What is the memory configuration difference?

A: The Radeon Pro 460 has 4 GB of dedicated GDDR5 memory on a 128 bit bus with 81.28 GB/s bandwidth. The Radeon 680M uses system shared memory, with bandwidth listed as System Dependent.

Q: Which GPU has higher clock speeds?

A: The Radeon 680M runs at a base clock of 2000 MHz and boosts to 2200 MHz. The Radeon Pro 460 has a base clock of 850 MHz and boosts to 907 MHz.

Q: Are there any tests where the Radeon Pro 460 wins?

A: In the head-to-head benchmark data, the Pro 460 does not win any shared tests. It does have a Geekbench Metal score of 20,426, but the Radeon 680M has no recorded Metal result for comparison.

The Verdict

The data supports a clear split based on workload. If you run OpenCL or Vulkan compute tasks, the Radeon 680M is the better part by a wide margin. Its 34.9% OpenCL lead and 23.4% Vulkan lead are too large to ignore. The 680M also offers ray tracing support, DirectX 12 Ultimate, and more than double the FP32 throughput. For modern gaming or GPU compute in Windows or Linux environments, the 680M is the practical choice.

The Radeon Pro 460 makes sense only in specific contexts. Its Geekbench Metal score of 20,426 suggests strong performance in Apple's Metal API, where the 680M has no recorded data. The Pro 460 also has dedicated GDDR5 memory, which avoids the system shared memory dependency of the 680M. If you need a GPU for a Mac environment that relies on Metal, or if you want predictable dedicated memory bandwidth without depending on system RAM, the Pro 460 remains viable.

The production status difference matters. The Pro 460 is end-of-life, released in 2016. The 680M is active, released in 2023, and has a defined successor in the Navi III IGP. For new system builds, the 680M is the forward-looking option. The Pro 460 only makes sense if you are working within an existing Mac platform that requires it.

Look at the rival positioning for context. The Pro 460's average score sits within 0.5% of the Radeon 780M and within 0.7% of the RTX 4060. The 680M sits within 0.5% of the RTX 3050 OEM and 0.7% above the RX 7600. In aggregate terms, the Pro 460 punches above its age, but the 680M delivers better performance in the specific APIs that matter for modern workloads. Choose the 680M for compute and gaming. Choose the Pro 460 only if you are locked into a Metal-centric Mac environment.