AMD Radeon 680M vs AMD Radeon 760M

Specifications of GPUs

GPU Comparison Result

Radeon 680M vs Radeon 760M: Graphics Are Nearly Tied, So CPU and Price Matter More

The Radeon 680M and Radeon 760M are AMD integrated GPUs for laptops and mini PCs. The 680M arrived in 2022 with Ryzen 6000 processors, using 12 RDNA 2 compute units and a 6 nm process. The 760M followed a year later with RDNA 3, a 4 nm process, and eight active compute units out of 12. Despite the newer architecture, real-world gaming performance is very similar.

In 3DMark Time Spy Graphics, both perform around the level of the older GeForce GTX 1050 Ti Max-Q. That is enough for online multiplayer games and older titles at 1080p, but demanding 2025 releases often require lower resolutions and upscaling even at low settings.

Both GPUs are relevant again in 2026 because AMD has reused them in renamed processors. The Radeon 680M appears in the Ryzen 7 170, while the 760M is used in the Ryzen 5 240, which is based on the Ryzen 5 8645HS. As a result, the CPU, the laptop itself, and the price matter more than the GPU name alone.

Key Differences

Specification Radeon 680M Radeon 760M
Architecture RDNA 2 RDNA 3
Compute units / stream processors 12 / 768 8 / 512
Clock speed 2.2-2.4 GHz 2.4-2.8 GHz
Theoretical FP32 About 3.4 TFLOPS Up to 5.7 TFLOPS
Process node 6 nm 4 nm
Memory support DDR5-4800, LPDDR5-6400 DDR5-5600, LPDDR5x-7500
Typical CPUs Ryzen 7 7735HS, Ryzen 7 170 Ryzen 5 7640HS, 8645HS, 240
AV1 Decode Decode and encode
Drivers Separate support branch Main driver branch
Debut 2022 2023

Position in AMD's Lineup

The Radeon 680M is the full 12-CU graphics configuration of the Rembrandt design. It appears in processors such as the Ryzen 7 6800H/HS/U, 7735HS, and the newer Ryzen 7 170. Six-core Ryzen 5 Rembrandt processors usually use the weaker Radeon 660M with six compute units.

The Radeon 760M is a cut-down version of the 780M. The Phoenix silicon physically contains 12 compute units, but only eight are enabled in these Ryzen 5 models. The 760M appears in the Ryzen 5 7640HS/U, 8640HS/U, 8645HS, and the newer Ryzen 5 230 and 240.

The Ryzen 7 170 is based on the older Zen 3+ architecture, while the Ryzen 5 240 uses Zen 4. The model numbers can therefore be misleading: the Ryzen 5 can be noticeably faster than the Ryzen 7 in single-threaded workloads.

RDNA 3's higher theoretical FP32 throughput does not translate directly into gaming performance. As a result, the 760M's eight compute units running at higher clock speeds mostly catch up with the 680M's 12 older units. More practical advantages of the 760M include AV1 encoding, faster memory support, and the main driver branch.

Graphics Benchmarks

Average results across laptops and mini PCs:

Benchmark Radeon 680M Radeon 760M Difference
3DMark Time Spy Graphics 2214 2260 760M +2%
3DMark Fire Strike Graphics 6672 6481 680M +3%
3DMark Steel Nomad 367 399 760M +9%
3DMark Speed Way 359 394 760M +10%
Wild Life Extreme Unlimited 3812 3382 680M +13%
Geekbench 6 GPU Vulkan 25,773 24,596 680M +5%

There is virtually no difference in Time Spy. The 760M looks slightly faster in newer benchmarks, but the sample size is small. Steel Nomad and Speed Way results for the 760M, for example, are based on only a few devices.

For comparison, the Radeon 780M averages around 2,659 points in Time Spy Graphics, roughly 18-20% higher than either GPU.

Results From Specific Devices

Device CPU and GPU Power Limits, W Time Spy Graphics
Lenovo IdeaPad Slim 5 15ARP10 Ryzen 7 7735HS, 680M 45/30 2430
Beelink SER6 Pro Ryzen 7 7735HS, 680M 45/38 2395
Lenovo ThinkPad E14 Gen 6 Ryzen 7 7735HS, 680M 35/30 2203
Lenovo Legion Go S Ryzen Z2 Go, 680M 40/30 2058
GMK NucBox K16 Ryzen 7 7735HS, 680M 50/50 1518
Lenovo Yoga 7 14AHP9 Ryzen 5 8640HS, 760M 41/31 2495
Acer Swift Edge 16 Ryzen 5 8640U, 760M 30/25 2293
GMKtec NucBox M6 Ultra Ryzen 5 7640HS, 760M N/A 2281
Lenovo LOQ 15APH8 Ryzen 5 7640HS, 760M N/A 2234
HP ZBook Firefly 14 G10 A Ryzen 5 PRO 7640HS, 760M 55/41 1997

The best 760M result in this table is only 3% higher than the best 680M result. The performance spread between devices using the same GPU is much larger.

The IdeaPad Slim 5 with the 680M clearly shows the effect of power limits. It scores 1,486 points at 15/12 W, 2,430 at 45/30 W, and 2,520 at 50/36 W. Dropping the sustained power limit to 12 W cuts performance by around 40%, while increasing it beyond 30 W adds very little.

This means the implementation in a specific laptop or mini PC can have a greater effect on performance than the choice between the 680M and 760M.

Gaming Performance

On the Radeon 760M at 1920×1080, F1 24 delivers around 58-66 FPS at low settings, Ghost of Tsushima around 44 FPS, Call of Duty: Black Ops 6 around 36-38 FPS, and Counter-Strike 2 about 105 FPS.

Demanding newer games are much harder to run. Battlefield 6 manages around 27 FPS at low settings, Kingdom Come: Deliverance II around 17-27 FPS, Oblivion Remastered about 14 FPS, Doom: The Dark Ages around 12 FPS, and Borderlands 4 about 10 FPS.

The Lenovo Legion Go S is a useful example for the Radeon 680M. In Cyberpunk 2077 at 1080p and low settings, it delivers 37.8 FPS at 30-35 W, 21.6 FPS at 15-20 W, and only 10.5 FPS at 7 W.

Overall, both GPUs are suitable for online multiplayer games and older titles at 1080p. Games from 2023-2024 often require low settings and upscaling, while the most demanding 2025 releases may require dropping to 720p.

Memory also matters. For gaming, 16 GB of dual-channel RAM should be considered the minimum, since the integrated GPU uses part of system memory as video memory.

Both GPUs have more than enough performance for office work, web browsing, video playback, and multiple displays. The 760M has one additional advantage: hardware AV1 encoding.

CPU Performance

The Radeon 680M is commonly paired with eight-core Ryzen 7 processors based on Zen 3+, while the 760M is usually paired with six-core Ryzen 5 chips based on Zen 4.

Benchmark Ryzen 7 7735HS / 170 Ryzen 5 7640HS / 8645HS / 240
Geekbench 6, single-core ~2085 ~2550-2580
Geekbench 6, multi-core ~9430 ~11,000
Cinebench 2024, single-core 89 104
Cinebench 2024, multi-core 671 654-698
Cinebench R23, single-core 1528 1742
Cinebench R23, multi-core 12,783 12,554-13,013

Zen 4 is around 14-24% faster in single-threaded performance. The gap is smaller in multi-threaded workloads because the Ryzen 7 7735HS has eight older cores compared with six Zen 4 cores.

At the same price, the Ryzen 5 240 or 8645HS is therefore faster in web browsers, office applications, and other workloads that depend heavily on single-threaded performance.

Drivers and FSR

Since late 2025, AMD has moved RDNA 1 and RDNA 2 GPUs to a separate support branch, while RDNA 3 remains on the main driver branch. The Radeon 680M continues to receive fixes and optimizations, but newer architectures have higher priority.

FSR 4.1 does not currently support either of these integrated GPUs. The Radeon 760M uses the newer RDNA 3 architecture and remains on AMD's main driver branch, while the RDNA 2-based 680M has moved to the separate support branch.

Where They Are Used and What They Cost

Device Configuration Graphics Price
Lenovo IdeaPad Slim 3 15ARP10 Ryzen 7 170, 16 GB, 512 GB 680M $500-550
Beelink SER5 MAX Ryzen 7 7735HS, 24 GB, 500 GB 680M Around $519
GMK NucBox K16 Ryzen 7 7735HS, 32 GB, 512 GB 680M $590-680
Lenovo Legion Go S Ryzen Z2 Go, 16 GB, 512 GB 680M $989-1049
GMKtec NucBox M6 Ultra Ryzen 5 7640HS, 32 GB, 512 GB-1 TB 760M $429-549
Beelink SER8 Ryzen 7 8845HS, 32 GB, 1 TB 780M $449-599

The biggest problem with 680M and 760M devices in 2026 is pricing. Mini PCs using these GPUs often cost as much as systems with the faster Radeon 780M.

Laptop pricing is more favorable. The IdeaPad Slim 3 with the Ryzen 7 170 has sold for $500-550 on sale. At that price, it offers eight CPU cores, 16 GB of memory, and Radeon 680M graphics. Paying the regular $850-1000 asking price for the same configuration makes little sense.

Older Radeon 760M laptops based on the Ryzen 5 8640HS and 8645HS make sense only at a substantial discount. In gaming laptops with a discrete GPU, the integrated 760M has little influence on the buying decision.

The Legion Go S with the Ryzen Z2 Go increased in price from $649 to $989-1049 in spring 2026. That is too expensive for a handheld with Radeon 680M-class graphics.

Alternatives

Integrated GPU Typical Processor Steel Nomad vs Radeon 760M
Radeon 840M Ryzen AI 5 340 208 -48%
Radeon 680M Ryzen 7 7735HS, 170 367 -8%
Radeon 760M Ryzen 5 7640HS, 8645HS, 240 399 0%
Radeon 780M Ryzen 7 8845HS, 260 453 +13%
Radeon 860M Ryzen AI 7 350 465 +16%
Radeon 890M Ryzen AI 9 HX 370 548 +37%

The Radeon 780M is the most important alternative. It is around 18% faster than the 760M in Time Spy Graphics and 13% faster in Steel Nomad. If a mini PC with the 780M costs the same or only slightly more, there is little reason to choose the 680M or 760M for gaming.

Despite belonging to a newer platform, the Radeon 840M is significantly slower. In Steel Nomad, it scores roughly half as much as the 760M. As a result, inexpensive Ryzen 7 170 laptops can still be faster in games than newer Ryzen AI 5 340 models.

The Radeon 860M and 890M are already noticeably faster than both the 680M and 760M.

Conclusion

In 2026, the Radeon 680M and 760M fall well behind today's faster integrated GPUs. Neither has a consistent performance advantage over the other: the 760M wins some benchmarks, the 680M wins others, and the difference is often smaller than the variation between individual devices.

Both are still suitable for office work, video playback, online multiplayer games, and older titles at 1080p. Demanding modern games require low settings, upscaling, or a lower resolution.

At the same price, a device with the Radeon 760M is the better choice. It is usually paired with a faster Zen 4 processor, supports AV1 encoding, and remains on AMD's main driver branch. A 680M-based device is worth considering when it is significantly cheaper or offers better features such as a superior display or longer battery life.

Reasonable prices as of September 2026:

  • A laptop with a Ryzen 7 170 or 7735HS and 16 GB of dual-channel memory - around $550 or less.
  • A mini PC with the 680M or 760M - only if it is at least $100 cheaper than a comparable Radeon 780M model, roughly $350-450 depending on memory and storage.
  • A Legion Go S with the Ryzen Z2 Go - around $600-650, not $989-1049.

Performance itself is not yet the main problem with the Radeon 680M and 760M. The real problem starts when a device using either GPU costs as much as a system with a Radeon 780M or newer integrated graphics.

Basic

Label Name
AMD
AMD
Launch Date
January 2022
January 2023
Platform
Integrated
Integrated
Model Name
Radeon 680M
Radeon 760M
Generation
Navi II IGP
Navi III IGP
Base Clock
2000MHz
1500MHz
Boost Clock
2200MHz
2800MHz
Shading Units
?
The most fundamental processing unit is the Streaming Processor (SP), where specific instructions and tasks are executed. GPUs perform parallel computing, which means multiple SPs work simultaneously to process tasks.
768
512
Transistors
13,100 million
-
RT Cores
12
8
Compute Units
12
8
TMUs
?
Texture Mapping Units (TMUs) serve as components of the GPU, which are capable of rotating, scaling, and distorting binary images, and then placing them as textures onto any plane of a given 3D model. This process is called texture mapping.
48
32
L1 Cache
128 KB per Array
128 KB per Array
L2 Cache
2MB
2MB
Bus Interface
PCIe 4.0 x8
-
Foundry
TSMC
TSMC
Process Size
6 nm
4 nm
Architecture
RDNA 2.0
RDNA 3.0
TDP
50W
-

Memory Specifications

Memory Size
System Shared
System Shared
Memory Type
System Shared
System Shared
Memory Bus
?
The memory bus width refers to the number of bits of data that the video memory can transfer within a single clock cycle. The larger the bus width, the greater the amount of data that can be transmitted instantaneously, making it one of the crucial parameters of video memory. The memory bandwidth is calculated as: Memory Bandwidth = Memory Frequency x Memory Bus Width / 8. Therefore, when the memory frequencies are similar, the memory bus width will determine the size of the memory bandwidth.
System Shared
System Shared
Memory Clock
SystemShared
SystemShared
Bandwidth
?
Memory bandwidth refers to the data transfer rate between the graphics chip and the video memory. It is measured in bytes per second, and the formula to calculate it is: memory bandwidth = working frequency × memory bus width / 8 bits.
System Dependent
System Dependent

Display and Media

Outputs
Portable Device Dependent
Portable Device Dependent

Theoretical Performance

Pixel Rate
?
Pixel fill rate refers to the number of pixels a graphics processing unit (GPU) can render per second, measured in MPixels/s (million pixels per second) or GPixels/s (billion pixels per second). It is the most commonly used metric to evaluate the pixel processing performance of a graphics card.
70.40 GPixel/s
44.80 GPixel/s
Texture Rate
?
Texture fill rate refers to the number of texture map elements (texels) that a GPU can map to pixels in a single second.
105.6 GTexel/s
89.60 GTexel/s
FP16 (half)
?
An important metric for measuring GPU performance is floating-point computing capability. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy.
6.758 TFLOPS
11.469 TFLOPS
FP64 (double)
?
An important metric for measuring GPU performance is floating-point computing capability. Double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy, while single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
211.2 GFLOPS
358.4 GFLOPS
FP32 (float)
?
An important metric for measuring GPU performance is floating-point computing capability. Single-precision floating-point numbers (32-bit) are used for common multimedia and graphics processing tasks, while double-precision floating-point numbers (64-bit) are required for scientific computing that demands a wide numeric range and high accuracy. Half-precision floating-point numbers (16-bit) are used for applications like machine learning, where lower precision is acceptable.
3.245 TFlops
5.734 TFlops

Miscellaneous

Vulkan Version
?
Vulkan is a cross-platform graphics and compute API by Khronos Group, offering high performance and low CPU overhead. It lets developers control the GPU directly, reduces rendering overhead, and supports multi-threading and multi-core processors.
1.3
1.3
OpenCL Version
2.0
2.1
OpenGL
4.6
4.6
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
Power Connectors
None
None
ROPs
?
The Raster Operations Pipeline (ROPs) is primarily responsible for handling lighting and reflection calculations in games, as well as managing effects like anti-aliasing (AA), high resolution, smoke, and fire. The more demanding the anti-aliasing and lighting effects in a game, the higher the performance requirements for the ROPs; otherwise, it may result in a sharp drop in frame rate.
32
16
Shader Model
6.7
6.7

Advantages

AMD Radeon 680M
Radeon 680M
  • More Shading Units: 768 (768 vs 512)
AMD Radeon 760M
Radeon 760M
  • Higher Boost Clock: 2800MHz (2200MHz vs 2800MHz)
  • Newer Launch Date: January 2023 (January 2022 vs January 2023)

FP32 (float)

Radeon 680M
3.245 TFlops
Radeon 760M
+77% 5.734 TFlops

3DMark Time Spy

Radeon 680M
+1% 2351
Radeon 760M
2329

Blender

Radeon 680M
+39% 254
Radeon 760M
183

SiliconCat Rating

697
Ranks 697 among all GPU on our website
496
Ranks 496 among all GPU on our website
Radeon 680M
Radeon 760M

Related GPU Comparisons