NVIDIA GeForce GTX 860M vs NVIDIA GeForce RTX 5080 SUPER Comparison

# NVIDIA GeForce RTX 5080 SUPER vs NVIDIA GeForce GTX 860M

The RTX 5080 SUPER is a desktop flagship-class GPU built on Blackwell 2.0, while the GTX 860M is a mobile Kepler part from 2014 that has long reached end-of-life status. The data shows a generational chasm: the RTX 5080 SUPER posts an average benchmark score of 3075 in 3DMark Steel Nomad DX12, whereas the GTX 860M averages 2967 across its suite of legacy tests. Despite the relatively close average scores, these two cards serve completely different purposes, and the RTX 5080 SUPER outperforms the GTX 860M by 3.6% in the nearest-rival comparison, with the 860M sitting at a -3.6% deltaPct relative to the 5080 SUPER.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The RTX 5080 SUPER averages 3075 in 3DMark Steel Nomad DX12, while the GTX 860M averages 2967 across its benchmark suite. The 5080 SUPER leads by 3.6% according to the deltaPct in the nearest-rival data.

Q: What are the biggest architectural differences between these two GPUs?

A: The RTX 5080 SUPER uses the GB203 chip on a 5 nm TSMC process with Blackwell 2.0 architecture, whereas the GTX 860M uses the GK104 chip on a 28 nm process with Kepler architecture. The 5080 SUPER has 45,600 million transistors versus 3,540 million, and it features 84 RT cores and 336 tensor cores, neither of which exist on the 860M.

Q: How do memory configurations compare?

A: The RTX 5080 SUPER has 24 GB of GDDR7 memory on a 256-bit bus with 1.02 TB/s bandwidth, while the GTX 860M has 2 GB of GDDR5 on a 128-bit bus with 80.00 GB/s bandwidth. That is a 12x difference in capacity and roughly 12.75x in bandwidth.

Q: Which GPU supports modern APIs?

A: The RTX 5080 SUPER supports DirectX 12 Ultimate (12_2), Vulkan 1.4, and OpenGL 4.6. The GTX 860M supports DirectX 12 (11_0), Vulkan 1.2.175, and OpenGL 4.6, meaning the 5080 SUPER has a higher DirectX feature level and a newer Vulkan version.

Q: What is the power consumption difference?

A: The RTX 5080 SUPER has a TDP of 415 W and requires a 16-pin power connector, while the GTX 860M has a TDP of 75 W and uses no external power connectors (it is an IGP-class mobile part).

Q: Are these GPUs in the same performance percentile?

A: Both sit at the 19th percentile versus all GPUs, though this reflects very different benchmark pools. The 5080 SUPER's single Steel Nomad score of 3075 places it near the Quadro P1000 (3163, -2.8%) and Intel Arc Pro B60 (3182, -3.4%), while the 860M's 2967 average places it near the GTX 750 Ti (2953, 0.5%) and RTX 4060 Ti 8 GB (2913, 1.9%).

Architecture Differences

The RTX 5080 SUPER is built on Blackwell 2.0 architecture using the GB203 chip fabricated on TSMC's 5 nm process. It packs 45,600 million transistors into a 378 mm² die, yielding a transistor density of 120.6M per mm². The GTX 860M, by contrast, uses the GK104 chip on TSMC's 28 nm process with 3,540 million transistors across a 294 mm² die, giving a density of just 12.0M per mm². The 5080 SUPER's density advantage is roughly 10x, reflecting over a decade of process improvement.

Compute resources diverge massively. The 5080 SUPER has 10,752 shading units, 336 TMUs, and 112 ROPs, while the 860M has 1,152 shading units, 96 TMUs, and 16 ROPs. The 5080 SUPER also adds 84 RT cores and 336 tensor cores, hardware the Kepler-based 860M entirely lacks. Pixel rate on the 5080 SUPER is 293.1 GPixel/s versus 21.96 GPixel/s on the 860M, and texture rate is 879.3 GTexel/s versus 87.84 GTexel/s.

Clock speeds tell a similar story. The 5080 SUPER runs at a 2295 MHz base and 2617 MHz boost, while the 860M operates at 797 MHz base and 915 MHz boost. The 5080 SUPER's FP32 throughput is 56.28 TFLOPS (with FP16 at 1:1), compared to just 2.108 TFLOPS on the 860M, which has no listed FP16 capability. The 860M supports DirectX 12 at the 11_0 feature level and Vulkan 1.2.175, whereas the 5080 SUPER reaches DirectX 12 Ultimate (12_2) and Vulkan 1.4. Both support OpenGL 4.6.

Memory architecture is another generation leap. The 5080 SUPER uses 24 GB of GDDR7 on a 256-bit bus with 1.02 TB/s bandwidth and a 2000 MHz base memory clock (32 Gbps effective). The 860M uses 2 GB of GDDR5 on a 128-bit bus with 80.00 GB/s bandwidth at 1250 MHz (5 Gbps effective). The 5080 SUPER connects via PCIe 5.0 x16, while the 860M uses PCIe 3.0 x16.

Head-to-Head Benchmarks

The head-to-head benchmark table is empty, so direct same-test comparisons are unavailable. Instead, the nearest-rival data provides the clearest cross-reference. The RTX 5080 SUPER's 3DMark Steel Nomad DX12 score of 3075 places it 3.6% ahead of the GTX 860M's average score of 2967. In the 5080 SUPER's rival set, the Quadro P1000 scores 3163 (-2.8% relative to the 5080 SUPER), the Intel Arc Pro B60 scores 3182 (-3.4%), and the GeForce 820A scores 2983 (+3.1%). The 860M's rival set shows the GTX 750 Ti at 2953 (+0.5% relative to the 860M), the GeForce 820A at 2983 (-0.5%), the Quadro P600 at 2923 (+1.5%), and the RTX 4060 Ti 8 GB at 2913 (+1.9%).

The 5080 SUPER's single benchmark result is a modern DX12 test, while the 860M's averages span Geekbench Metal (4900), OpenCL (10461), Vulkan (9648), and Passmark tests ranging from DirectX 9 (55) to DirectX 12 (13) to G3D (3081) and GPU Compute (1251). The 860M's strongest Passmark result is DirectX 9 at 55, and its G3D score of 3081 is close to the 5080 SUPER's Steel Nomad score. However, these are different workloads from different eras, so direct numerical comparison is misleading. The deltaPct of 3.6% favoring the 5080 SUPER is the most reliable cross-GPU signal available.

The Verdict

The RTX 5080 SUPER is the clear choice for anyone building a modern desktop system that needs current API support, ray tracing hardware, tensor cores, and massive memory bandwidth. Its 24 GB GDDR7 frame buffer, 1.02 TB/s bandwidth, and 56.28 TFLOPS FP32 performance make it suitable for demanding DX12 Ultimate workloads, and its 84 RT cores and 336 tensor cores enable features the GTX 860M cannot handle at all. The 415 W TDP and 16-pin connector indicate a high-end desktop part, and the 999 USD launch MSRP positions it as a flagship product.

The GTX 860M, by contrast, is an end-of-life mobile GPU with a 75 W TDP, no external power connector, and portable-device-dependent display outputs. Its 2 GB GDDR5 and 80.00 GB/s bandwidth are severely limited by modern standards, and its Kepler architecture lacks ray tracing and tensor cores entirely. Its DirectX 12 support is capped at the 11_0 feature level, and Vulkan support stops at 1.2.175. It makes sense only for legacy systems or very light workloads where the 75 W power draw and IGP form factor are advantages.

The data does not support choosing the 860M for any modern use case. The 5080 SUPER wins on every meaningful architectural metric: process node (5 nm vs 28 nm), transistor count (45,600 vs 3,540 million), shading units (10,752 vs 1,152), memory capacity (24 GB vs 2 GB), bandwidth (1.02 TB/s vs 80.00 GB/s), and FP32 throughput (56.28 vs 2.108 TFLOPS). The only areas where the 860M is less demanding are power (75 W vs 415 W) and physical footprint (IGP vs dual-slot, 304 mm length).

Specification Differences

| Specification | RTX 5080 SUPER | GTX 860M |

|---|---|---|

| Architecture | Blackwell 2.0 | Kepler |

| Chip | GB203 | GK104 |

| Process Node | 5 nm | 28 nm |

| Transistors | 45,600 million | 3,540 million |

| Die Size | 378 mm² | 294 mm² |

| Transistor Density | 120.6M / mm² | 12.0M / mm² |

| Base Clock | 2295 MHz | 797 MHz |

| Boost Clock | 2617 MHz | 915 MHz |

| Memory Size | 24 GB | 2 GB |

| Memory Type | GDDR7 | GDDR5 |

| Memory Bus | 256 bit | 128 bit |

| Memory Bandwidth | 1.02 TB/s | 80.00 GB/s |

| Shading Units | 10752 | 1152 |

| TMUs | 336 | 96 |

| ROPs | 112 | 16 |

| RT Cores | 84 | None |

| Tensor Cores | 336 | None |

| Pixel Rate | 293.1 GPixel/s | 21.96 GPixel/s |

| Texture Rate | 879.3 GTexel/s | 87.84 GTexel/s |

| FP32 | 56.28 TFLOPS | 2.108 TFLOPS |

| FP16 | 56.28 TFLOPS (1:1) | None |

| TDP | 415 W | 75 W |

| Slot Width | Dual-slot | IGP |

| Power Connectors | 1x 16-pin | None |

| Bus Interface | PCIe 5.0 x16 | PCIe 3.0 x16 |

| Display Outputs | 1x HDMI 2.1b, 3x DisplayPort 2.1b | Portable Device Dependent |

| DirectX | 12 Ultimate (12_2) | 12 (11_0) |

| Vulkan | 1.4 | 1.2.175 |

| Production Status | Active | End-of-life |

| Release Date | 2025-12-31 | 2014-03-09 |

Where Each One Wins

The RTX 5080 SUPER wins in every compute and rendering category that matters for modern gaming and professional workloads. Its 56.28 TFLOPS FP32 and 56.28 TFLOPS FP16 (1:1) throughput dominate the 860M's 2.108 TFLOPS FP32, and its 84 RT cores and 336 tensor cores enable ray-traced rendering and AI-accelerated features that the Kepler architecture cannot execute. The 24 GB GDDR7 frame buffer at 1.02 TB/s bandwidth supports high-resolution textures and large datasets, while the 860M's 2 GB GDDR5 at 80.00 GB/s is a hard bottleneck. The 5080 SUPER's DirectX 12 Ultimate and Vulkan 1.4 support ensure compatibility with current and near-future titles, whereas the 860M's DirectX 12 (11_0) and Vulkan 1.2.175 limit it to older or less demanding software.

The GTX 860M wins only on power efficiency and physical flexibility. Its 75 W TDP versus the 5080 SUPER's 415 W means it can run in thin-and-light laptops without external power, and its IGP form factor with no dedicated power connector makes it a drop-in mobile solution. The 860M also has a wider benchmark footprint, with results across Geekbench Metal, OpenCL, Vulkan, and multiple Passmark DX tests, but those scores are all modest. For a builder considering a desktop upgrade, the 5080 SUPER is the only rational pick; the 860M is a legacy part that the data shows is outclassed in every performance metric. The 3.6% average-score advantage for the 5080 SUPER understates the real gap, because the 5080 SUPER's single modern DX12 test is far more demanding than the 860M's legacy suite.