NVIDIA GeForce GTX 860M

The NVIDIA GeForce GTX 860M is a mobile graphics processor from the Kepler generation, built on a 28 nm process at TSMC. It is positioned as an end-of-life product, having been released in early 2014, and its benchmark data reflects a mixed legacy, with a strong showing in some compute-oriented tests but a modest overall standing in the current GPU landscape.

Benchmark Performance

The GTX 860M’s performance profile is defined by a clear split between legacy DirectX workloads and modern compute APIs. In the PassMark suite, the highest score recorded is for DirectX 9 at 55 points, while DirectX 11 drops to 23 points, and DirectX 10 falls further to 15 points. The lowest API-specific result is DirectX 12 at 13 points, indicating that the Kepler architecture struggles significantly with modern graphics pipelines. This pattern is consistent with the GPU’s core design, which lacks dedicated hardware for newer features, relying instead on software fallbacks that penalize performance.

The compute-oriented benchmarks tell a different story. In Geekbench OpenCL, the GTX 860M achieves a score of 10,472, and in Geekbench Vulkan, it reaches 9,551. The Geekbench Metal score is notably lower at 4,900, suggesting that Apple’s Metal API is not a strength for this GPU. The PassMark GPU Compute score of 1,251 further illustrates that raw compute throughput, while not exceptional, is comparatively stronger than the graphics-specific results. The average benchmark score across all tests is 2,959, which places the GPU at the 17th percentile of all GPUs — a figure that underscores its position as a low-to-mid-range part in the current market.

The deltaPct values against nearest rivals reveal a narrow competitive band. The GTX 860M trails the NVIDIA GeForce GT 730M by just 0.5%, and it is 0.8% behind the NVIDIA GeForce 820A. Conversely, it leads the NVIDIA GeForce RTX 4060 Ti 8 GB and the NVIDIA GeForce GT 555M by 1.6% each. These margins are remarkably tight, suggesting that in the aggregate, the GTX 860M performs nearly identically to a much older entry-level part (GT 730M) and a modern mid-range card (RTX 4060 Ti), at least in terms of the average score metric. The data indicates that architectural generation matters less than raw benchmark aggregation when comparing across such disparate product lines.

How It Compares

Against the NVIDIA GeForce GT 730M, the GTX 860M is essentially a statistical tie, with a delta of -0.5% meaning the GT 730M holds a marginal lead. This is surprising given the GTX 860M’s higher transistor count of 3,540 million and larger die size of 294 mm², but the average scores of 2,959 and 2,975 respectively show that real-world performance convergence is possible despite hardware differences.

The NVIDIA GeForce 820A presents a similar picture, with the GTX 860M trailing by 0.8%. The 820A’s average score of 2,983 is the highest among the listed rivals, yet the gap is within a rounding error. This suggests that for legacy DirectX and compute tasks, the GTX 860M offers no meaningful advantage over this entry-level part, which is a critical observation for anyone evaluating its long-term viability.

The comparison with the NVIDIA GeForce RTX 4060 Ti 8 GB is the most counterintuitive. The RTX 4060 Ti, a modern card with a far more recent architecture, posts an average score of 2,913, which is 1.6% lower than the GTX 860M’s average. This does not imply the GTX 860M is faster in modern games; rather, it reflects that the aggregate benchmark suite is heavily weighted toward older APIs where Kepler’s raw shader count can still compete, while the RTX 4060 Ti’s advantages in ray tracing and DirectX 12 are not captured in this particular average.

Finally, the NVIDIA GeForce GT 555M is the closest rival in terms of score, with a delta of 1.6% in favor of the GTX 860M. The GT 555M’s average score of 2,913 is identical to the RTX 4060 Ti’s, which highlights how benchmark aggregation can flatten the differences between generations. The GTX 860M’s 1152 shading units and 96 texture mapping units provide a solid base, but the 16 ROPs are a bottleneck that limits fill-rate-heavy workloads.

Ray Tracing and Feature Set

The GTX 860M has no dedicated ray tracing cores and no tensor cores, as these are absent from the fact pack. Its architecture is Kepler, which predates the hardware-accelerated ray tracing found in later NVIDIA generations. Consequently, any ray tracing workload would have to be processed through the standard shading units, resulting in severe performance degradation. The API support confirms this limitation: the GPU supports DirectX 12, but only at feature level 11_0, meaning it cannot fully utilize the DirectX 12 Ultimate feature set that includes ray tracing and mesh shaders.

On the software side, the GTX 860M supports OpenGL 4.6 and Vulkan 1.2.175. The Vulkan support is noteworthy because it allows access to modern compute and graphics features on this older hardware, which partially explains the relatively high Geekbench Vulkan score of 9,551. However, the lack of hardware-accelerated ray tracing means that any game relying on that feature will not run natively at acceptable speeds. The GPU’s pixel rate is 21.96 GPixel/s and its texture rate is 87.84 GTexel/s, which are modest figures that cap its ability to handle high-resolution textures and complex lighting effects.

Who Should Consider It

Benchmark results indicate that the GTX 860M is best suited for legacy gaming at lower resolutions and with reduced graphical settings. The PassMark DirectX 9 score of 55 is the strongest graphics result, suggesting that older titles from that era will run relatively smoothly. For DirectX 11 games, the score drops to 23, meaning that 1080p gaming with medium-to-high settings would likely push the GPU to its limits, and users should expect frame rates below 60 fps in demanding scenes.

At higher resolutions, such as 1440p or 4K, the GTX 860M is not a viable option. The 80.00 GB/s memory bandwidth and 128-bit bus width are insufficient for the data throughput required at those resolutions, and the 16 ROPs will struggle with pixel fill rates. The GPU’s 17th percentile standing among all GPUs reinforces that it is a low-tier part, appropriate only for users who play older games or are willing to accept low settings in modern titles. For any current AAA game, the DirectX 12 score of 13 is a clear warning that performance will be poor.

Memory Subsystem

The GTX 860M is equipped with 2 GB of GDDR5 memory, which was standard for its time but is now considered minimal for modern gaming. The memory bus is 128 bits wide, and the memory clock runs at 1250 MHz, translating to 5 Gbps effective and a bandwidth of 80.00 GB/s. This bandwidth is a critical constraint: it is roughly half of what mid-range desktop GPUs offered in the same era, and it directly impacts performance at higher resolutions and with anti-aliasing enabled.

For high-resolution textures, the 2 GB capacity is a limiting factor. Many modern games at 1080p require more than 2 GB of VRAM for high-quality texture packs, leading to texture swapping and stuttering. The 80.00 GB/s bandwidth further compounds this issue, as the GPU cannot move data quickly enough to keep the shading units fed. The pixel rate of 21.96 GPixel/s, derived from the 16 ROPs and the boost clock of 915 MHz, means that fill-rate-heavy effects like bloom and motion blur will cause noticeable frame drops.

Power and Cooling

The GTX 860M has a TDP of 75 W, which is modest for a mobile GPU and reflects its Kepler architecture’s efficiency on the 28 nm process. The fact pack lists no suggested PSU, but the power connector is listed as “None,” indicating that this is an integrated GPU (IGP) soldered onto the motherboard, not a discrete card with external power requirements. This makes it suitable for laptops where thermal and power constraints are strict, but it also means that users cannot upgrade or replace the GPU.

The slot width is listed as “IGP,” and the display outputs are described as “Portable Device Dependent,” confirming that this is not a desktop part. The lack of a power connector simplifies installation in a laptop chassis, but it also caps the GPU’s ability to boost beyond its 915 MHz limit under sustained load, as thermal headroom is typically limited in thin-and-light designs. The base clock of 797 MHz and boost clock of 915 MHz are relatively close, indicating that the GPU runs near its maximum frequency most of the time.

FAQ

Q: What is the average benchmark score of the GTX 860M?

A: The average benchmark score across all tests is 2,959, placing it at the 17th percentile of all GPUs.

Q: How does the GTX 860M compare to the NVIDIA GeForce RTX 4060 Ti 8 GB?

A: The GTX 860M is 1.6% faster in average benchmark score, with a score of 2,959 versus 2,913 for the RTX 4060 Ti 8 GB.

Q: Does the GTX 860M support hardware ray tracing?

A: No, the GPU has no ray tracing cores and no tensor cores, and its DirectX 12 support is limited to feature level 11_0.

Q: What is the memory bandwidth of the GTX 860M?

A: The memory bandwidth is 80.00 GB/s, achieved through a 128-bit bus and GDDR5 memory running at 1250 MHz (5 Gbps effective).

Q: What is the TDP of the GTX 860M, and does it require a power connector?

A: The TDP is 75 W, and the power connector is listed as “None,” indicating it is an integrated part for portable devices.

Q: In which benchmark does the GTX 860M score the highest?

A: The highest score is in Geekbench OpenCL at 10,472, followed by Geekbench Vulkan at 9,551.