AMD Radeon 760M vs AMD Radeon 860M Comparison

FAQ

Q: Which GPU has the higher average benchmark score?

A: The AMD Radeon 860M records an average benchmark score of 26,401, while the AMD Radeon 760M averages 6,019. This places the 860M in the 72nd percentile of all GPUs, compared to the 760M's 35th percentile.

Q: How do the two compare in Geekbench OpenCL?

A: The Radeon 860M scores 22,759 in Geekbench OpenCL, which is 11% higher than the Radeon 760M's 20,255. This is one of the two recorded head-to-head benchmark results.

Q: Which GPU wins in Geekbench Vulkan?

A: The Radeon 760M takes the Vulkan test with a score of 30,336 versus the Radeon 860M's 30,043, a margin of 1%. This gives each GPU one win in the head-to-head comparison.

Q: What are the nearest rivals for each GPU based on average score?

A: The Radeon 760M sits within 0.5% of the AMD Radeon RX 6400 (6,001), NVIDIA GeForce GTX 770M (6,000), NVIDIA RTX PRO 6000 Blackwell Server (5,996), and NVIDIA Quadro P2000 (6,049). The Radeon 860M is within 1.1% of the NVIDIA GeForce MX550 (26,421), NVIDIA GeForce RTX 5060 (26,331), AMD Radeon RX 5700 XT 50th Anniversary (26,553), and NVIDIA RTX A4000 (26,683).

Q: Do both GPUs use the same process node?

A: Yes, both are fabricated by TSMC on a 4 nm process. However, the Radeon 760M is based on the Phoenix chip with RDNA 3.0 architecture, while the Radeon 860M uses the Krackan Point chip with RDNA 3.5 architecture.

Q: What is the release date difference between the two?

A: The Radeon 760M was released on January 30, 2024, and the Radeon 860M followed on February 28, 2025. Both are currently listed as Active in production status.

Architecture Differences

The AMD Radeon 760M and AMD Radeon 860M share the same fundamental compute layout but diverge in architectural generation. Both integrate 512 shading units, 32 texture mapping units, 16 ROPs, and 8 ray tracing cores, yet the 760M uses RDNA 3.0 architecture on the Phoenix chip, while the 860M moves to RDNA 3.5 on the Krackan Point chip. This architectural refresh is the primary differentiator, as it brings changes to instruction scheduling and resource management without altering the raw unit counts.

The process node remains identical at 4 nm from TSMC, and the transistor count for the 760M is listed at 25,390 million across a 178 mm² die, giving a density of 142.6 million transistors per mm². The 860M's transistor count and die size are recorded as unknown, so a direct physical comparison is not possible from the database. Both GPUs are integrated graphics parts (IGP) with a 15 W TDP, and neither requires external power connectors.

Clock behavior differs notably between the two. The 760M has a base clock of 800 MHz and a boost clock of 2,599 MHz, while the 860M starts at a lower 600 MHz base but boosts to 3,000 MHz. The higher boost clock on the 860M drives its pixel rate to 48.00 GPixel/s and texture rate to 96.00 GTexel/s, compared to 41.58 GPixel/s and 83.17 GTexel/s on the 760M. Interestingly, the FP32 compute figures show the opposite relationship: the 760M delivers 5.323 TFLOPS, while the 860M is rated at 3.072 TFLOPS, despite the higher clocks on the latter. This suggests the RDNA 3.5 implementation uses a different FP32 execution path per clock cycle.

Memory subsystems are identical in configuration: both use System Shared memory with system-dependent bandwidth and a PCIe 4.0 x8 bus interface. API support matches across DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Display outputs differ only in their dependency: the 760M is motherboard-dependent, while the 860M is portable device dependent, reflecting their target platforms.

Head-to-Head Benchmarks

The recorded head-to-head data contains two benchmark results, and each GPU claims one victory. In Geekbench OpenCL, the Radeon 860M scores 22,759 against the Radeon 760M's 20,255, a delta of 11% in favor of the newer architecture. This is the larger margin of the two tests and indicates a meaningful compute advantage for the 860M in OpenCL workloads.

The Geekbench Vulkan test flips the result. The Radeon 760M scores 30,336, edging out the Radeon 860M's 30,043 by 1%. The margin is narrow, but it shows that the older RDNA 3.0 design retains a slight edge in Vulkan API performance, likely due to its higher FP32 throughput. The 760M's 5.323 TFLOPS versus the 860M's 3.072 TFLOPS aligns with this outcome, as Vulkan compute paths often leverage FP32 arithmetic directly.

Beyond the direct comparison, the average benchmark scores reveal a broader gap. The 860M averages 26,401 across its recorded tests, while the 760M averages 6,019. This 4.4x difference is driven by the 860M's superior OpenCL result and the fact that the 760M's average includes several lower-scoring legacy tests (PassMark DirectX 9 at 65, DirectX 10 at 19, DirectX 11 at 52, DirectX 12 at 25, G2D at 890, G3D at 5,310, and GPU Compute at 2,840), which pull its mean down. The 860M has no recorded PassMark results in the database, so its average relies solely on the two Geekbench scores.

The nearest rival data also places the two in different competitive brackets. The 760M sits among older discrete GPUs like the NVIDIA Quadro P2000 (6,049, delta -0.5%) and the NVIDIA GeForce GTX 770M (6,000, delta 0.3%), while the 860M competes with more modern parts including the NVIDIA GeForce RTX 5060 (26,331, delta 0.3%) and the AMD Radeon RX 5700 XT 50th Anniversary (26,553, delta -0.6%). This separation in rival scores confirms that the 860M operates in a substantially higher performance class despite identical core counts and TDP.

Specification Differences

The two GPUs differ across several recorded specifications, while sharing others. The architecture is RDNA 3.0 for the 760M and RDNA 3.5 for the 860M. The chip names differ: Phoenix for the 760M versus Krackan Point for the 860M. The generation field also varies, with the 760M listed under "Navi III IGP (Phoenix)" and the 860M under "Navi III IGP (Strix Point Mobile)".

Clock speeds show the most notable divergence. The 760M runs at 800 MHz base and 2,599 MHz boost, while the 860M runs at 600 MHz base and 3,000 MHz boost. This results in different pixel and texture rates: 41.58 GPixel/s and 83.17 GTexel/s for the 760M, versus 48.00 GPixel/s and 96.00 GTexel/s for the 860M. The FP32 and FP16 compute figures are identical within each part (1:1 ratio) but differ across them: 5.323 TFLOPS for the 760M and 3.072 TFLOPS for the 860M.

Transistor count and die size are listed for the 760M (25,390 million transistors, 178 mm², 142.6M / mm²) but are unknown for the 860M. The release dates differ by roughly 13 months: January 30, 2024 for the 760M and February 28, 2025 for the 860M. Display outputs are motherboard-dependent for the 760M and portable device dependent for the 860M.

Shared specifications include the 4 nm TSMC process, 512 shading units, 32 TMUs, 16 ROPs, 8 RT cores, 15 W TDP, PCIe 4.0 x8 interface, System Shared memory configuration, DirectX 12 Ultimate (12_2), OpenGL 4.6, Vulkan 1.4, and the absence of power connectors. Both are IGP parts with no launch MSRP in the database.

Where Each One Wins

The Radeon 760M wins in scenarios that favor raw FP32 compute and Vulkan API performance. Its 5.323 TFLOPS is 73% higher than the 860M's 3.072 TFLOPS, and this translates directly into its 1% Vulkan victory (30,336 versus 30,043). Applications that rely heavily on Vulkan rendering paths, including certain game engines and compute workloads, will show better results on the 760M. Its higher base clock (800 MHz versus 600 MHz) also suggests better sustained performance at lower utilization states, though the boost clock is lower at 2,599 MHz versus 3,000 MHz.

The Radeon 860M wins in OpenCL compute and overall average performance. Its 22,759 OpenCL score beats the 760M by 11%, and its average benchmark score of 26,401 is more than four times higher than the 760M's 6,019. The 860M's higher boost clock of 3,000 MHz and superior pixel and texture rates (48.00 GPixel/s and 96.00 GTexel/s) give it an edge in fill-rate-bound tasks. The percentile placement confirms this: the 860M sits at the 72nd percentile of all GPUs, while the 760M is at the 35th percentile. For users running OpenCL-accelerated software, video encoding, or general compute workloads, the 860M is the stronger choice.

The nearest rival data reinforces this split. The 760M's rivals include the NVIDIA Quadro P2000 and GTX 770M, which are older professional and mobile parts with modest performance. The 860M's rivals include the NVIDIA GeForce RTX 5060 and RTX A4000, which are modern discrete GPUs with substantially higher performance ceilings. This indicates that the 860M should be considered in a different tier entirely, despite sharing the same core configuration and power envelope.

In terms of use-case split, the 760M is better suited to Vulkan-centric gaming on lightweight platforms where its higher FP32 throughput can be utilized, and where the motherboard-dependent display outputs indicate desktop or standard laptop integration. The 860M is better suited to portable devices (per its display output dependency) and applications that leverage OpenCL, where its architectural refinements in RDNA 3.5 and higher boost clock deliver measurable gains. The 860M's release date of February 2025 also means it benefits from a newer driver and software ecosystem relative to the January 2024 launch of the 760M, though the database does not quantify driver impacts directly.