NVIDIA GeForce GTX 1650 vs NVIDIA GeForce GTX 950M Comparison

The NVIDIA GeForce GTX 950M and the NVIDIA GeForce GTX 1650 are separated by four years and a full architectural leap, yet both target the entry-level mobile and desktop segments. The GTX 1650 is decisively faster in every benchmark recorded, but the GTX 950M remains a relevant baseline for older titles and legacy workloads. The data shows that the GTX 1650 wins 2 out of 2 head-to-head tests, with the GTX 950M failing to secure a single victory. For anyone choosing between the two today, the GTX 1650 is the only sensible pick for modern gaming, while the GTX 950M serves as a historical reference point or a low-cost fallback for undemanding applications.

The Verdict

The GTX 1650 is the clear performance winner, and the benchmark data leaves no room for ambiguity. In Geekbench OpenCL, the GTX 1650 scores 29,629 against the GTX 950M’s 9,745, a delta of -67.1% from the GTX 1650's perspective—meaning the GTX 950M trails by roughly two-thirds. In Geekbench Vulkan, the gap widens further: the GTX 1650 posts 33,042 versus 6,525, a -80.3% delta. These are not marginal improvements; they represent a generational chasm.

The GTX 950M’s average benchmark score is 8,135, sitting at the 42nd percentile of all GPUs, while the GTX 1650’s average is actually lower at 7,472 (40th percentile), despite its dominant head-to-head wins. This paradox stems from the GTX 1650’s broader test suite, which includes multiple Passmark DX9–DX12 and compute tests where it scores modestly (e.g., 39 in Passmark DX10, 124 in DX9). The GTX 950M only has two Geekbench entries, skewing its average upward. Thus, the average scores are not directly comparable; the head-to-head tests are the reliable indicator.

For a user with a GTX 950M laptop, the upgrade to a GTX 1650 is transformative. For a buyer of a new budget system, the GTX 1650’s 75 W TDP and lack of power connectors make it a drop-in replacement for older Maxwell cards, but its dual-slot cooler and 229 mm length require case clearance. The GTX 950M is end-of-life, while the GTX 1650 is also end-of-life but newer (released 2019-04-22 versus 2015-03-12). The verdict is simple: pick the GTX 1650 for any gaming or compute task; reserve the GTX 950M for legacy compatibility or as a paperweight.

Architecture Differences

The GTX 950M is built on the GM107 chip using NVIDIA's Maxwell architecture, fabricated on a 28 nm TSMC process. It packs 1,870 million transistors into a 148 mm² die, yielding a transistor density of 12.6 million per mm². The GTX 1650, in contrast, uses the TU117 chip with the Turing architecture on a 12 nm TSMC node, housing 4,700 million transistors in a 200 mm² die—a density of 23.5 million per mm². This is a 2.5× increase in transistor count and nearly double the density, explaining the massive performance gap.

Core configurations differ substantially. The GTX 950M has 640 shading units, 40 texture mapping units (TMUs), and 16 raster operations pipelines (ROPs). The GTX 1650 nearly doubles the TMUs to 56 and ROPs to 32, while increasing shading units to 896. Pixel rate jumps from 17.98 GPixel/s to 53.28 GPixel/s, and texture rate from 44.96 GTexel/s to 93.24 GTexel/s. FP32 compute scales from 1,438.7 GFLOPS to 2.984 TFLOPS, with the GTX 1650 also offering FP16 at 5.967 TFLOPS (2:1 rate) that the GTX 950M lacks entirely.

Memory subsystems tell a similar story. Both have 4 GB, but the GTX 950M uses DDR3 at 900 MHz with 1,800 Mbps effective speed, yielding 28.80 GB/s bandwidth on a 128-bit bus. The GTX 1650 uses GDDR5 at 2,001 MHz with 8 Gbps effective speed, hitting 128.1 GB/s—over 4× the bandwidth. Clock speeds also rise: the GTX 950M runs at 993 MHz base and 1,124 MHz boost, while the GTX 1650 starts at 1,485 MHz and boosts to 1,665 MHz. Both are 75 W TDP, but the GTX 950M is an IGP (integrated graphics package) with no power connectors, while the GTX 1650 is dual-slot with a suggested 250 W PSU.

API support differs in DirectX: the GTX 950M supports DirectX 12 (11_0) while the GTX 1650 supports DirectX 12 (12_1). Both offer OpenGL 4.6 and Vulkan 1.4. The GTX 1650 also has PCIe 3.0 x16, doubling the GTX 950M’s x8 interface. The GTX 1650’s display outputs include 1x DVI, 1x HDMI 2.0, and 1x DisplayPort 1.4a, whereas the GTX 950M is "Portable Device Dependent." Notably, the GTX 1650 has no RT or tensor cores, despite being Turing, so ray tracing is absent on both.

Where Each One Wins

The GTX 1650 wins every single benchmark category in the head-to-head data. In Geekbench OpenCL, it scores 29,629 versus 9,745, a 204% advantage. In Geekbench Vulkan, it scores 33,042 versus 6,525, a 406% advantage. The GTX 1650 also wins the broader Passmark suite, but those tests lack GTX 950M scores for comparison, so they only establish the GTX 1650's absolute capabilities: 7,880 in G3D, 3,048 in GPU Compute, and 561 in G2D.

The GTX 950M’s only "win" is in the nearest-rival comparison. Its average score of 8,135 places it 0.1% above the AMD Radeon R9 M360 (8,129) and 0.4% above the NVIDIA GeForce 945M (8,099). It also edges the NVIDIA GRID K2 by 0.7%. Against the GTX 980 (8,167), it trails by 0.4%. This suggests the GTX 950M is competitive within its own 2015-era peer group, but that group is obsolete.

For use cases, the GTX 1650 is the choice for any modern DirectX 12 (12_1) title, Vulkan-based games, or compute workloads leveraging FP16. The GTX 950M, with DirectX 12 (11_0) and no FP16, is limited to older DirectX 11 games or OpenGL applications. The GTX 1650’s 128.1 GB/s bandwidth is critical for texture-heavy scenes, while the GTX 950M’s 28.80 GB/s will bottleneck even modest resolutions. The GTX 1650 also supports HDMI 2.0 and DisplayPort 1.4a for modern monitors, whereas the GTX 950M relies on laptop-specific outputs.

In practical terms, the GTX 950M suits 720p gaming at low settings for titles from its era (pre-2016). The GTX 1650 can handle 1080p medium-to-high settings for many games, given its 2.984 TFLOPS and 53.28 GPixel/s fill rate. The GTX 1650’s dual-slot design means it needs a proper desktop chassis, while the GTX 950M was often soldered into laptops. Neither has RT or tensor cores, so no ray tracing or DLSS.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The GTX 950M has an average benchmark score of 8,135, which is higher than the GTX 1650’s 7,472. However, this is misleading because the GTX 950M only has two Geekbench results, while the GTX 1650 includes ten tests, several of which (Passmark DX9, DX10, DX11, DX12) score low due to legacy API overhead.

Q: What is the biggest performance gap between the two cards?

A: The largest gap is in Geekbench Vulkan, where the GTX 1650 scores 33,042 against the GTX 950M’s 6,525, a delta of -80.3% from the GTX 1650’s perspective. This means the GTX 950M achieves less than 20% of the GTX 1650’s Vulkan performance.

Q: Do both cards support the same DirectX version?

A: No. The GTX 950M supports DirectX 12 (11_0), while the GTX 1650 supports DirectX 12 (12_1). Both support OpenGL 4.6 and Vulkan 1.4, but the higher DirectX feature level on the GTX 1650 enables newer rendering techniques.

Q: Is the GTX 1650 more power-hungry than the GTX 950M?

A: Both have a 75 W TDP, but the GTX 1650 suggests a 250 W PSU and uses a dual-slot cooler with no power connectors. The GTX 950M is an IGP with no power connectors and a single-slot integrated design, making it more suited for compact laptops.

Q: How does each card compare to its nearest rivals?

A: The GTX 950M sits 0.1% above the AMD Radeon R9 M360 and 0.4% above the NVIDIA GeForce 945M, but 0.4% below the GTX 980. The GTX 1650 is 0.3% above the AMD Radeon HD 8850M and 0.4% above the Intel UHD Graphics 750, but 1% below the Intel Arc A310.

Q: Which card has more memory bandwidth?

A: The GTX 1650 has 128.1 GB/s from GDDR5 memory, while the GTX 950M has 28.80 GB/s from DDR3. This is a 4.45× advantage for the GTX 1650, which directly impacts texture streaming and high-resolution performance.

Head-to-Head Benchmarks

The head-to-head data includes exactly two tests, and the GTX 1650 wins both decisively. In Geekbench OpenCL, the GTX 1650 scores 29,629 versus the GTX 950M’s 9,745. The delta is -67.1% when measured from the GTX 1650’s score, meaning the GTX 950M is only 32.9% as fast. This test stresses general-purpose compute on the GPU, and the GTX 1650’s 896 shading units and 2.984 TFLOPS FP32 power dwarf the GTX 950M’s 640 units and 1,438.7 GFLOPS. The GTX 1650’s FP16 capability (5.967 TFLOPS) is absent on the GTX 950M, further widening the compute gap.

The second test, Geekbench Vulkan, is even more lopsided. The GTX 1650 achieves 33,042, while the GTX 950M manages only 6,525—a delta of -80.3%. Vulkan is a low-overhead API that benefits from raw throughput, and the GTX 1650’s 53.28 GPixel/s pixel rate and 93.24 GTexel/s texture rate are 3× and 2.07× higher than the GTX 950M’s 17.98 GPixel/s and 44.96 GTexel/s, respectively. The GTX 1650 also has double the ROPs (32 versus 16), which helps with fill-rate-bound scenes.

Outside the direct head-to-head, the GTX 1650’s Passmark scores provide context for its strengths. Its G3D score of 7,880 and GPU Compute of 3,048 show balanced raster and compute performance. The GTX 950M has no Passmark entries, so only the Geekbench results are directly comparable. In both, the GTX 1650’s advantage is overwhelming, driven by a 12 nm process versus 28 nm, GDDR5 versus DDR3, and a 2.5× transistor budget. The GTX 950M’s nearest-rival data shows it is competitive only with 2015-era midrange parts like the R9 M360 and 945M, which share its 28 nm Maxwell DNA. Against the GTX 1650, the older card cannot compete in any measurable metric.