AMD Radeon 760M vs NVIDIA GeForce GTX 675M Comparison

Head-to-Head Benchmarks

The only directly comparable benchmark in the database is Geekbench OpenCL, and the results are decisive. The AMD Radeon 760M scores 20,255, while the NVIDIA GeForce GTX 675M scores 6,946. This gives the AMD part a 65.7% advantage in this test, a substantial margin that reflects the generational gap between the two designs.

In the database's overall scoring, the NVIDIA GeForce GTX 675M averages 6,946 points, placing it in the 39th percentile of all GPUs. Its nearest rivals show how tightly clustered this performance tier is: the AMD Radeon R5 M240 scores 6,975 (-0.4% difference), the NVIDIA GeForce GTX 680M scores 7,023 (-1.1%), the NVIDIA T600 scores 7,035 (-1.3%), and the AMD FirePro M5100 scores 6,830 (+1.7%). The GTX 675M sits squarely within this group, with no rival exceeding a 1.7% delta in either direction. This indicates that the old Fermi-based mobile part remains competitive with its direct contemporaries, even if it cannot match the newer RDNA 3 part.

The AMD Radeon 760M, by contrast, averages 6,019 points across all its recorded benchmarks, placing it in the 35th percentile of all GPUs. Its nearest rivals cluster around a slightly lower performance envelope: the AMD Radeon RX 6400 scores 6,001 (+0.3%), the NVIDIA GeForce GTX 770M scores 6,000 (+0.3%), the NVIDIA RTX PRO 6000 Blackwell Server scores 5,996 (+0.4%), and the NVIDIA Quadro P2000 scores 6,049 (-0.5%). This suggests the 760M's average is pulled down by its diverse benchmark suite, which includes some low-scoring DirectX 9 and DirectX 10 results. The OpenCL result, however, tells a different story: the 760M is more than three times faster than the GTX 675M in that specific compute workload.

Looking at the broader benchmark suite for the Radeon 760M, the data reveals a heterogeneous performance profile. In 3DMark Steel Nomad DX12, it scores 400. In Geekbench Vulkan, it scores 30,336. Passmark results vary widely: 19 in DirectX 10, 52 in DirectX 11, 25 in DirectX 12, 65 in DirectX 9, 890 in G2D, 5,310 in G3D, and 2,840 in GPU Compute. The GTX 675M has only the single Geekbench OpenCL result recorded, so cross-test comparisons are limited. Still, the OpenCL gap alone is enough to establish the AMD part as the compute winner in this pairing.

FAQ

Q: Which GPU wins the head-to-head benchmark?

A: The AMD Radeon 760M wins, with a Geekbench OpenCL score of 20,255 versus 6,946 for the NVIDIA GeForce GTX 675M, a 65.7% delta in favor of AMD.

Q: How does the GTX 675M compare to its closest rivals?

A: The GTX 675M averages 6,946 points. The AMD Radeon R5 M240 is 0.4% behind at 6,975, the GTX 680M is 1.1% behind at 7,023, the NVIDIA T600 is 1.3% behind at 7,035, and the AMD FirePro M5100 is 1.7% ahead at 6,830. All deltas are within a narrow 3% band.

Q: What does the Radeon 760M's average score indicate?

A: Its average of 6,019 points across all benchmarks places it in the 35th percentile. The closest rival, the NVIDIA Quadro P2000, is 0.5% ahead, while the AMD Radeon RX 6400, NVIDIA GeForce GTX 770M, and NVIDIA RTX PRO 6000 Blackwell Server are all within 0.4% or less.

Q: Is the Radeon 760M consistently faster across all tests?

A: No. Its Passmark results vary significantly: 65 in DirectX 9, 52 in DirectX 11, 25 in DirectX 12, 19 in DirectX 10, and 5,310 in G3D. The OpenCL and Vulkan scores are far higher at 20,255 and 30,336 respectively, showing strong compute performance but weaker legacy API performance.

Q: How do the pixel and texture rates compare?

A: The Radeon 760M achieves 41.58 GPixel/s and 83.17 GTexel/s, while the GTX 675M achieves 9.92 GPixel/s and 39.68 GTexel/s. The AMD part is roughly four times faster in pixel throughput and about twice as fast in texture throughput.

Q: What is the production status of each GPU?

A: The NVIDIA GeForce GTX 675M is end-of-life, released in 2012. The AMD Radeon 760M is active, released in 2024.

Architecture Differences

The NVIDIA GeForce GTX 675M uses the GF114 chip built on the Fermi 2.0 architecture, fabricated on a 40 nm process at TSMC. It integrates 1,950 million transistors on a 332 mm² die, yielding a transistor density of 5.9 million per square millimeter. The design includes 384 shading units, 64 texture mapping units, and 32 ROPs. It offers DirectX 12 (11_0) support, OpenGL 4.6, and has no Vulkan support recorded. The architecture predates ray tracing and tensor core hardware entirely.

The AMD Radeon 760M uses the Phoenix chip built on RDNA 3.0, fabricated on a 4 nm process, also at TSMC. It integrates 25,390 million transistors on a 178 mm² die, yielding a much higher transistor density of 142.6 million per square millimeter. The design includes 512 shading units, 32 texture mapping units, 16 ROPs, and 8 ray tracing cores. It supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The FP16 compute rate matches FP32 at 5.323 TFLOPS (1:1 ratio), indicating full-rate half-precision support, which the Fermi architecture lacks entirely.

The process node difference is stark: 40 nm versus 4 nm. This allows the Radeon 760M to pack over 13 times more transistors into a smaller die, which directly explains its higher compute throughput despite a significantly lower power envelope. The GTX 675M's Fermi architecture lacks modern features like hardware ray tracing, which the Radeon 760M provides with its 8 RT cores.

The memory subsystems are fundamentally different. The GTX 675M uses dedicated 2 GB GDDR5 memory on a 256-bit bus with 96.00 GB/s bandwidth. The Radeon 760M uses system shared memory, with bandwidth listed as system dependent. This means the AMD part's memory performance is contingent on the host system's RAM configuration, a key consideration for gaming workloads.

Specification Differences

The process node differs substantially: the GTX 675M is on 40 nm, while the Radeon 760M is on 4 nm. Transistor count favors AMD massively: 25,390 million versus 1,950 million. Die size goes the other way: the NVIDIA chip is 332 mm², the AMD chip is 178 mm². Transistor density is 5.9M / mm² for NVIDIA versus 142.6M / mm² for AMD.

Clock speeds are recorded differently. The GTX 675M has no base or boost clock listed, only a memory clock of 750 MHz (3 Gbps effective). The Radeon 760M has a base clock of 800 MHz and a boost clock of 2,599 MHz. The memory configuration is also different: the GTX 675M uses 2 GB GDDR5 on a 256-bit bus with 96.00 GB/s bandwidth, while the Radeon 760M uses system shared memory with system dependent bandwidth.

Compute resources differ: the NVIDIA part has 384 shading units, 64 TMUs, and 32 ROPs; the AMD part has 512 shading units, 32 TMUs, and 16 ROPs. The AMD part adds 8 ray tracing cores, which the NVIDIA part lacks entirely. Pixel rate is 9.92 GPixel/s for NVIDIA versus 41.58 GPixel/s for AMD. Texture rate is 39.68 GTexel/s versus 83.17 GTexel/s. FP32 compute is 952.3 GFLOPS for NVIDIA versus 5.323 TFLOPS for AMD. The AMD part also lists FP16 at 5.323 TFLOPS (1:1), which the NVIDIA part does not record.

Power consumption differs dramatically: the GTX 675M has a TDP of 100 W, while the Radeon 760M has a TDP of 15 W. The form factor also differs: the NVIDIA part is an MXM Module with MXM-B (3.0) bus interface, while the AMD part is an IGP with PCIe 4.0 x8 interface. Display outputs are portable device dependent for NVIDIA and motherboard dependent for AMD. DirectX support favors AMD with 12 Ultimate (12_2) versus 12 (11_0) for NVIDIA. Vulkan support is present on AMD (1.4) but not recorded on NVIDIA. Production status is end-of-life for NVIDIA and active for AMD.

Where Each One Wins

The AMD Radeon 760M wins decisively in compute throughput. Its Geekbench OpenCL score of 20,255 dwarfs the GTX 675M's 6,946, and its FP32 output of 5.323 TFLOPS is more than five times the NVIDIA part's 952.3 GFLOPS. The 760M also wins in pixel and texture throughput, with 41.58 GPixel/s versus 9.92 GPixel/s and 83.17 GTexel/s versus 39.68 GTexel/s respectively. For modern workloads that leverage DirectX 12 Ultimate features, ray tracing, or Vulkan, the Radeon 760M is the only option of the two, as the GTX 675M lacks ray tracing cores and Vulkan support entirely.

The Radeon 760M also wins on power efficiency. Its 15 W TDP is a fraction of the GTX 675M's 100 W, yet it delivers significantly higher performance in every measured compute metric. This makes it the clear choice for battery-powered or thermally constrained systems where the integrated form factor is an advantage.

The NVIDIA GeForce GTX 675M has no recorded benchmark wins in this pairing. Its only advantage lies in its dedicated memory subsystem: 2 GB of GDDR5 on a 256-bit bus with 96.00 GB/s bandwidth provides predictable, consistent memory performance that does not depend on system RAM. The Radeon 760M's system shared memory is system dependent, so its effective bandwidth varies with the host platform. In scenarios where a system has slow or insufficient shared memory, the GTX 675M's dedicated VRAM could provide more stable performance in memory-intensive tasks.

The GTX 675M also has a higher ROP count (32 versus 16) and more TMUs (64 versus 32), which could theoretically benefit certain fill-rate-bound or texture-heavy workloads, though the recorded pixel and texture rates show the AMD part achieving far higher absolute throughput due to its clock speed advantage. The GTX 675M sits in the 39th percentile versus the 760M's 35th, indicating that in the broader GPU landscape, the older NVIDIA part's single benchmark places it slightly higher relative to all GPUs, but this is a narrow distinction driven by the 760M's lower-scoring legacy API tests.

For legacy DirectX 9 and DirectX 10 workloads, the GTX 675M's Fermi architecture may have an edge in compatibility, but the database shows the 760M scoring 65 in DirectX 9 and 19 in DirectX 10 on Passmark, which are low absolute numbers. The GTX 675M has no recorded results in these tests, so a direct comparison is not possible. Ultimately, the Radeon 760M is the stronger choice for any modern compute or graphics workload, while the GTX 675M retains relevance only in scenarios requiring dedicated VRAM with fixed bandwidth.