AMD Radeon RX 6400 vs NVIDIA GeForce GTX 670M Comparison

# Head-to-Head Benchmarks

The single available head-to-head data point is Geekbench OpenCL, and the result is decisive. The AMD Radeon RX 6400 scores 32,011 versus the NVIDIA GeForce GTX 670M’s 6,513. That is a delta of -79.7% from the perspective of the GTX 670M, meaning the RX 6400 is approximately 4.9 times faster in this compute workload. This is not a close contest; the margin is so large that it represents multiple generational leaps in raw throughput.

The RX 6400’s average benchmark score across all recorded tests is 6,001, which places it at the 35th percentile of all GPUs. The GTX 670M’s average is 6,513, putting it at the 38th percentile. Interestingly, the GTX 670M has a higher average score than the RX 6400, despite losing the OpenCL test by a wide margin. This is because the RX 6400’s average is dragged down by its Passmark DirectX 9 score of 93, its Passmark DirectX 10 score of 54, and its Passmark DirectX 11 score of 70. The GTX 670M has no corresponding low scores in the data, so its single OpenCL result stands as its entire benchmark profile.

The RX 6400’s nearest rivals in the database include the NVIDIA GeForce GTX 770M at 6,000 (0% delta), the NVIDIA RTX PRO 6000 Blackwell Server at 5,996 (0.1% delta), and the AMD FirePro W4100 at 5,987 (0.2% delta). The GTX 670M’s nearest rivals are the NVIDIA GeForce GT 555M at 6,493 (0.3% delta), the NVIDIA Quadro M5000M at 6,481 (0.5% delta), and the AMD Radeon Vega 10 Mobile at 6,476 (0.6% delta). The fact that the RX 6400 sits alongside professional workstation and server-class parts in its rival list, while the GTX 670M sits alongside older mobile parts, underscores the architectural gulf between them.

# Architecture Differences

The two GPUs come from different eras and different design philosophies. The GTX 670M is built on NVIDIA’s Fermi 2.0 architecture, using the GF114 chip fabricated on a 40 nm process at TSMC. The RX 6400 uses AMD’s RDNA 2.0 architecture, with the Navi 24 chip on a 6 nm process, also at TSMC. The process node difference is stark: 40 nm versus 6 nm. This explains the dramatic difference in transistor density. The GTX 670M packs 1,950 million transistors into a 332 mm² die, yielding a density of 5.9 million transistors per square millimeter. The RX 6400 crams 5,400 million transistors into just 107 mm², achieving 50.5 million transistors per square millimeter. That is an order of magnitude higher density.

The memory subsystems are also fundamentally different. The GTX 670M uses 1,536 MB of GDDR5 on a 192-bit bus, with memory clocked at 750 MHz (3 Gbps effective) and bandwidth of 72.00 GB/s. The RX 6400 uses 4 GB of GDDR6 on a 64-bit bus, with memory at 2,000 MHz (16 Gbps effective) and bandwidth of 128.0 GB/s. Despite having one-third the bus width, the RX 6400 delivers nearly double the bandwidth due to its much faster memory technology.

Compute resources differ substantially. The GTX 670M has 336 shading units, 56 texture mapping units, and 24 ROPs. The RX 6400 has 768 shading units, 48 TMUs, and 32 ROPs. The RX 6400 has more than double the shading units and more ROPs, though fewer TMUs. The pixel rate of the RX 6400 is 74.27 GPixel/s versus 8.372 GPixel/s for the GTX 670M. Texture rate is 111.4 GTexel/s versus 33.49 GTexel/s. Floating-point performance (FP32) is 3.565 TFLOPS versus 803.7 GFLOPS. The RX 6400 also has 12 ray tracing cores and supports FP16 at 7.130 TFLOPS (2:1 ratio), neither of which exist on the GTX 670M.

The RX 6400 supports DirectX 12 Ultimate (12_2), while the GTX 670M is limited to DirectX 12 (11_0). Both support OpenGL 4.6. The RX 6400 supports Vulkan 1.4; the GTX 670M has no Vulkan entry. The RX 6400 uses a PCIe 4.0 x4 interface, while the GTX 670M uses an MXM-B (3.0) bus. Power consumption favors the newer card: the RX 6400 has a TDP of 53 W, while the GTX 670M draws 75 W. The RX 6400 is single-slot with no power connectors and a suggested PSU of 250 W. The GTX 670M is an MXM module with no power connectors and portable-device-dependent display outputs. The RX 6400 outputs via 1x HDMI 2.1 and 1x DisplayPort 1.4a.

# The Verdict

The data is unambiguous in raw compute terms. The RX 6400 outperforms the GTX 670M by 79.7% in the only direct head-to-head benchmark available. Its FP32 throughput of 3.565 TFLOPS is over four times the GTX 670M’s 803.7 GFLOPS. Its pixel rate is nearly nine times higher, and its texture rate is more than triple. The RX 6400 also has twice the memory capacity, and its 128.0 GB/s bandwidth is 78% higher despite a 64-bit bus versus 192-bit.

However, the GTX 670M holds a higher percentile ranking (38th versus 35th) and a higher average benchmark score (6,513 versus 6,001). This is a statistical artifact of the RX 6400’s low DirectX 9, 10, and 11 Passmark scores, which are not directly comparable to the GTX 670M’s single OpenCL result. The GTX 670M’s nearest rivals are all within 0.6% of its score, suggesting it is a stable, mid-pack performer. The RX 6400’s nearest rivals include a workstation GPU and a server GPU, indicating it competes in a broader performance envelope.

For users prioritizing modern API support, ray tracing, and raw compute performance, the RX 6400 is the clear choice. For users constrained to legacy DirectX 9-11 workloads or who need a mobile MXM form factor, the GTX 670M retains relevance. The RX 6400’s launch MSRP was 159 USD. The GTX 670M has no recorded launch MSRP. The GTX 670M is end-of-life, as is the RX 6400. The GTX 670M was released in 2012, the RX 6400 in 2022. The RX 6400’s successor is Navi III; the GTX 670M’s successor is GeForce 700M.

# FAQ

Q: Which GPU has the higher Geekbench OpenCL score?

A: The AMD Radeon RX 6400 scores 32,011, while the NVIDIA GeForce GTX 670M scores 6,513. The RX 6400 wins by 79.7%.

Q: How does memory bandwidth compare between the two?

A: The RX 6400 has 128.0 GB/s bandwidth using 4 GB of GDDR6 on a 64-bit bus. The GTX 670M has 72.00 GB/s using 1,536 MB of GDDR5 on a 192-bit bus.

Q: What is the process node difference?

A: The GTX 670M uses a 40 nm process with 1,950 million transistors on a 332 mm² die. The RX 6400 uses a 6 nm process with 5,400 million transistors on a 107 mm² die.

Q: Does the RX 6400 support ray tracing?

A: Yes, the RX 6400 has 12 ray tracing cores and supports DirectX 12 Ultimate (12_2). The GTX 670M has no ray tracing cores and supports only DirectX 12 (11_0).

Q: Which GPU has a higher TDP?

A: The GTX 670M has a TDP of 75 W. The RX 6400 has a TDP of 53 W, making it the more power-efficient option.

Q: What is the average benchmark score for each GPU?

A: The GTX 670M has an average benchmark score of 6,513. The RX 6400 has an average of 6,001.

# Where Each One Wins

The RX 6400 wins in every scenario that stresses modern compute features. Its FP32 throughput of 3.565 TFLOPS versus 803.7 GFLOPS makes it superior for general-purpose compute, machine learning inference, and any workload that leverages Shader Model 6.0+ features. Its 12 ray tracing cores enable hardware-accelerated ray tracing, which the GTX 670M cannot do at all. The RX 6400’s 128.0 GB/s memory bandwidth and 4 GB frame buffer provide headroom for higher-resolution textures and larger datasets, while the GTX 670M’s 1,536 MB and 72.00 GB/s will bottleneck in such scenarios. The RX 6400’s Vulkan 1.4 support and DirectX 12 Ultimate (12_2) API compliance make it the only choice for titles that require these modern APIs.

The GTX 670M wins in one narrow but real category: legacy DirectX 9, 10, and 11 workloads. Its Passmark DirectX 9 score of 93, DirectX 10 score of 54, and DirectX 11 score of 70 are not directly comparable to the RX 6400’s scores of 93, 54, and 70 respectively—wait, the data actually shows the RX 6400 achieves exactly those same scores in those three Passmark tests. That means the two GPUs are effectively tied in legacy DirectX performance, with the GTX 670M having no numerical advantage in any recorded test. The GTX 670M’s MXM-B (3.0) form factor makes it the only option for certain laptop and portable systems where the RX 6400’s PCIe 4.0 x4 interface cannot be accommodated.

The RX 6400 also wins on efficiency. Its 53 W TDP versus 75 W means it delivers dramatically higher performance while consuming 29% less power. The RX 6400’s single-slot design with no external power connectors makes it easier to install in constrained chassis. The GTX 670M’s portable-device-dependent display outputs limit its connectivity options, while the RX 6400 offers standard HDMI 2.1 and DisplayPort 1.4a outputs.

In the single direct benchmark comparison, the RX 6400 has one win and the GTX 670M has zero wins. The RX 6400’s percentile ranking of 35th places it just below the GTX 670M’s 38th, but the RX 6400’s average score is dragged down by its low Passmark DirectX scores in legacy APIs. For users looking forward, the RX 6400 is the only sensible choice based on the data. For users locked into legacy systems with MXM slots, the GTX 670M remains functional but offers no performance advantage anywhere in the recorded benchmarks.