NVIDIA GeForce RTX 4070 vs NVIDIA GeForce RTX 5090 Mobile Comparison

NVIDIA
GEFORCE

NVIDIA GeForce RTX 4070

CORE STATE AD104
VRAM 12 GB
CLOCK SPEED 2475 MHz
TDP 200 W
BUS WIDTH 192 bit
ARCHITECTURE Ada Lovelace
nm
PROCESS 5 nm
LAUNCH DATE 2023
VS
NVIDIA
GEFORCE

GeForce RTX 5090 Mobile

CORE STATE GB203
VRAM 24 GB
CLOCK SPEED 1515 MHz
TDP 95 W
BUS WIDTH 256 bit
ARCHITECTURE Blackwell 2.0
nm
PROCESS 5 nm
LAUNCH DATE 2025

PERFORMANCE BENCHMARKS

3dmark_3dmark_steel_nomad_dx12
3,854
5,871
geekbench_opencl
154,858
201,834
geekbench_vulkan
174,152
198,405
passmark_directx_10
139
183
passmark_directx_11
244
269
passmark_directx_12
103
138
passmark_directx_9
320
324
passmark_g2d
1,164
1,057
passmark_g3d
26,927
30,034
passmark_gpu_compute
14,720
13,401

Analysis: NVIDIA GeForce RTX 4070 vs NVIDIA GeForce RTX 5090 Mobile

Head-to-Head Benchmarks

The recorded data shows a decisive overall victory for the NVIDIA GeForce RTX 5090 Mobile, which wins 8 of the 10 benchmark comparisons. The largest margin comes in the DirectX 12 3DMark Steel Nomad test, where the 5090 Mobile scores 5871 against the RTX 4070's 3854, a 52.3% advantage. This is the single biggest performance gap in the dataset and signals a substantial generational leap in modern API workloads.

The GeForce RTX 5090 Mobile also demonstrates strong superiority in compute-oriented tests. In Geekbench OpenCL, it scores 201834 versus 154858, a 30.3% lead, while in Geekbench Vulkan it posts 198405 against 174152, a 13.9% advantage. PassMark DirectX 10 shows a 31.7% gap (183 vs 139), and PassMark DirectX 12 reveals a 34% gap (138 vs 103). These results indicate that the 5090 Mobile's architectural improvements translate broadly across both graphics and general-purpose compute APIs.

The DirectX 11 test shows a more modest 10.2% lead (269 vs 244), while PassMark G3D, which aggregates overall 3D performance, lands at 11.5% ahead (30034 vs 26927). The legacy DirectX 9 test is nearly a tie, with the 5090 Mobile ahead by only 1.3% (324 vs 320), suggesting that older API workloads do not benefit from the newer architecture's features to the same degree.

However, the RTX 4070 claims two wins, both in specific categories. In PassMark G2D (2D graphics), the 4070 scores 1164 versus 1057, a 9.2% advantage. In PassMark GPU Compute, the 4070 posts 14720 against 13401, a 9% lead. The compute result is particularly notable because it contradicts the Geekbench OpenCL outcome, indicating that the 4070's higher boost clock and different scheduling characteristics favor certain compute workloads, while the 5090 Mobile's larger shader count dominates in others.

Looking at the broader database context, the 5090 Mobile holds an 84th percentile ranking among all GPUs, with an average benchmark score of 45152. Its closest rival in the database is the AMD Radeon Pro 5500 XT at 45384 (a -0.5% delta, meaning the 5090 Mobile is effectively at parity), followed by the NVIDIA GeForce RTX 4070 Ti at 44795 (0.8% ahead), the Intel Arc A730M at 45592 (1% behind), and the NVIDIA RTX 5880 Ada Generation at 45972 (1.8% behind). The RTX 4070, by contrast, sits at the 81st percentile with an average score of 37648, placing it near the NVIDIA Tesla P4 (37628, 0.1% delta) and the AMD Radeon RX Vega 56 (37507, 0.4% delta). The 5090 Mobile's average score is 20% higher than the 4070's, which aligns with the head-to-head deltas observed across most individual tests.

Architecture Differences

The two GPUs represent distinct architectural generations from NVIDIA. The RTX 5090 Mobile is built on the Blackwell 2.0 architecture using the GB203 chip, fabricated on a 5 nm process at TSMC with 45,600 million transistors on a 378 mm² die, yielding a transistor density of 120.6 million per mm². The RTX 4070 uses the Ada Lovelace architecture with the AD104 chip, also on a 5 nm TSMC process, but with 35,800 million transistors on a 294 mm² die, for a density of 121.8 million per mm². The density figures are nearly identical, but the 5090 Mobile packs roughly 27% more transistors into a 29% larger die.

The core configuration differences are substantial. The 5090 Mobile features 10,496 shading units, 328 texture mapping units, 112 ROPs, 82 RT cores, and 328 tensor cores. The RTX 4070 offers 5,888 shading units, 184 TMUs, 64 ROPs, 46 RT cores, and 184 tensor cores. This means the 5090 Mobile has 78% more shading units, 78% more TMUs, 75% more ROPs, 78% more RT cores, and 78% more tensor cores. The near-uniform percentage increases across these units suggest a scaling approach that preserves the architectural balance of the design.

Clock speeds tell a different story. The RTX 4070 operates at a base clock of 1920 MHz and a boost of 2475 MHz, while the 5090 Mobile runs at a substantially lower 990 MHz base and 1515 MHz boost. Despite the lower clocks, the 5090 Mobile achieves a higher FP32 throughput of 31.80 TFLOPS versus 29.15 TFLOPS for the 4070, thanks to its larger shader count. Pixel rate favors the 5090 Mobile at 169.7 GPixel/s versus 158.4 GPixel/s, while texture rate is nearly identical at 496.9 GTexel/s versus 455.4 GTexel/s. Both GPUs offer FP16 at a 1:1 ratio with FP32.

Memory architecture differs significantly. The 5090 Mobile uses 24 GB of GDDR7 on a 256-bit bus, delivering 896.0 GB/s of bandwidth. The RTX 4070 uses 12 GB of GDDR6X on a 192-bit bus, providing 504.2 GB/s. The 5090 Mobile thus offers double the capacity and 78% more bandwidth, which is critical for high-resolution textures and large datasets. Memory clock is also higher on the 5090 Mobile at 1750 MHz (28 Gbps effective) versus 1313 MHz (21 Gbps effective).

Power and physical design diverge considerably. The 5090 Mobile is rated at 95 W TDP and is an integrated GPU (IGP) with no dedicated power connectors, while the RTX 4070 draws 200 W TDP, is dual-slot, and requires a single 16-pin connector with a suggested 550 W PSU. The 5090 Mobile uses a PCIe 5.0 x16 interface, while the 4070 uses PCIe 4.0 x16. Display outputs on the 5090 Mobile are portable device dependent, whereas the 4070 offers 1x HDMI 2.1 and 3x DisplayPort 1.4a. Both support DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. The 5090 Mobile has a release date in March 2025 and is currently active, while the 4070 launched in April 2023 and is marked end-of-life.

FAQ

Q: Which GPU has a higher average benchmark score in the database?

A: The NVIDIA GeForce RTX 5090 Mobile has an average benchmark score of 45152, which is 20% higher than the RTX 4070's average of 37648. The 5090 Mobile also holds an 84th percentile ranking versus the 4070's 81st percentile.

Q: How much faster is the RTX 5090 Mobile in DirectX 12 workloads?

A: In the 3DMark Steel Nomad DirectX 12 test, the RTX 5090 Mobile scores 5871 versus the RTX 4070's 3854, a 52.3% advantage. In PassMark DirectX 12, the 5090 Mobile leads by 34% (138 vs 103).

Q: Are there any benchmarks where the RTX 4070 beats the RTX 5090 Mobile?

A: Yes, the RTX 4070 wins in two tests. It scores 1164 in PassMark G2D versus 1057 for the 5090 Mobile (a 9.2% lead), and it scores 14720 in PassMark GPU Compute versus 13401 (a 9% lead).

Q: What are the memory capacity and type differences between the two cards?

A: The RTX 5090 Mobile has 24 GB of GDDR7 memory on a 256-bit bus with 896.0 GB/s bandwidth. The RTX 4070 has 12 GB of GDDR6X on a 192-bit bus with 504.2 GB/s bandwidth.

Q: How do the power requirements compare?

A: The RTX 5090 Mobile has a 95 W TDP and is an integrated GPU with no power connectors. The RTX 4070 has a 200 W TDP, is dual-slot, requires a single 16-pin connector, and has a suggested PSU of 550 W.

Q: Which GPU has a higher boost clock?

A: The RTX 4070 has a much higher boost clock at 2475 MHz, compared to the RTX 5090 Mobile's 1515 MHz. Despite this, the 5090 Mobile produces higher FP32 throughput (31.80 TFLOPS vs 29.15 TFLOPS) due to its larger shader count.

Specification Differences

The two GPUs differ across nearly every major specification category. The RTX 5090 Mobile uses the GB203 chip on Blackwell 2.0 architecture, while the RTX 4070 uses the AD104 chip on Ada Lovelace. The 5090 Mobile is built with 45,600 million transistors on a 378 mm² die, versus 35,800 million on a 294 mm² die for the 4070. Transistor density is nearly equal at 120.6M/mm² versus 121.8M/mm², respectively.

Core counts are substantially different: the 5090 Mobile has 10,496 shading units, 328 TMUs, 112 ROPs, 82 RT cores, and 328 tensor cores, while the 4070 has 5,888 shading units, 184 TMUs, 64 ROPs, 46 RT cores, and 184 tensor cores. Clock speeds are inverted, with the 4070 running much faster at 1920 MHz base and 2475 MHz boost, versus 990 MHz base and 1515 MHz boost for the 5090 Mobile. The 5090 Mobile still edges ahead in FP32 (31.80 vs 29.15 TFLOPS), pixel rate (169.7 vs 158.4 GPixel/s), and texture rate (496.9 vs 455.4 GTexel/s).

Memory specifications show the 5090 Mobile with 24 GB GDDR7, 256-bit bus, 896.0 GB/s bandwidth, and 1750 MHz memory clock, versus the 4070's 12 GB GDDR6X, 192-bit bus, 504.2 GB/s bandwidth, and 1313 MHz memory clock. The 5090 Mobile has a 95 W TDP, IGP form factor, no power connectors, and PCIe 5.0 x16, while the 4070 has a 200 W TDP, dual-slot form factor, one 16-pin connector, suggested 550 W PSU, and PCIe 4.0 x16. The 4070 has physical dimensions of 240 mm length, 110 mm height, and 40 mm width, while the 5090 Mobile has no listed dimensions. The 5090 Mobile's display outputs are portable device dependent, whereas the 4070 has 1x HDMI 2.1 and 3x DisplayPort 1.4a. The 5090 Mobile released in March 2025 and is active, while the 4070 released in April 2023 and is end-of-life, with a launch MSRP of 599 USD.

The Verdict

The data supports a clear split decision based on workload. The NVIDIA GeForce RTX 5090 Mobile is the dominant choice for modern, API-heavy gaming and general compute tasks. Its 52.3% lead in DirectX 12 Steel Nomad, 30.3% lead in OpenCL, and 13.9% lead in Vulkan make it the superior option for current-generation game engines and GPU-accelerated applications. The 5090 Mobile also offers double the memory capacity (24 GB vs 12 GB) and 78% more bandwidth, which matters for large textures, high resolutions, and data-intensive workloads. Its lower TDP of 95 W versus 200 W is an additional advantage in mobile environments, despite the lower clock speeds.

The NVIDIA GeForce RTX 4070 retains relevance in two specific areas. Its 9.2% advantage in PassMark G2D suggests better 2D rendering performance, which may matter for desktop productivity or legacy applications. Its 9% lead in PassMark GPU Compute indicates that certain compute kernels favor the 4070's higher boost clock and architectural scheduling, even though OpenCL results lean heavily toward the 5090 Mobile. The 4070's physical card form factor and discrete display outputs also make it more appropriate for desktop builds, while the 5090 Mobile is designed as an integrated solution.

For users prioritizing raw modern gaming performance, ray tracing capabilities (82 RT cores versus 46), and memory capacity, the RTX 5090 Mobile is the clear choice from the recorded measurements. For users with legacy 2D-dependent workflows or specific compute patterns that favor the 4070's clock characteristics, the RTX 4070 remains a viable option, but the overall benchmark data shows the 5090 Mobile as the stronger performer in the vast majority of tests. The RTX 4070's end-of-life status and the 5090 Mobile's active production status further reinforce the generational shift, with the 5090 Mobile representing the current direction of NVIDIA's mobile GPU lineup.

DETAILED SPECIFICATIONS

SPECIFICATION
RTX 4070
RTX 5090 Mobile
Core Specs
Shading Units
5,888
10,496 +78.3%
Shaders
5,888
10,496 +78.3%
TMUs
184
328 +78.3%
ROPs
64
112 +75.0%
SM Count
46
82 +78.3%
Clocks
Base Clock
1920 MHz
990 MHz
Boost Clock
2475 MHz
1515 MHz
Memory Clock
1313 MHz 21 Gbps effective
1750 MHz 28 Gbps effective
Memory
Memory Size
12 GB
24 GB
VRAM (MB)
12,288
24,576 +100.0%
Memory Type
GDDR6X
GDDR7
Memory Bus
192 bit
256 bit
Bandwidth
504.2 GB/s
896.0 GB/s
Cache
L1 Cache
128 KB (per SM)
128 KB (per SM)
L2 Cache
36 MB
64 MB
Performance
Pixel Rate
158.4 GPixel/s
169.7 GPixel/s
Texture Rate
455.4 GTexel/s
496.9 GTexel/s
FP32 (TFLOPS)
29.15 TFLOPS
31.80 TFLOPS
FP64 (TFLOPS)
455.4 GFLOPS (1:64)
496.9 GFLOPS (1:64)
FP16 (TFLOPS)
29.15 TFLOPS (1:1)
31.80 TFLOPS (1:1)
AI/RT
RT Cores
46
82 +78.3%
Tensor Cores
184
328 +78.3%
Power
TDP
200 W
95 W
TDP (W)
200
95 -52.5%
Suggested PSU
550 W
Power Connectors
1x 16-pin
None
Architecture
Architecture
Ada Lovelace
Blackwell 2.0
GPU Name
AD104
GB203
Generation
GeForce 40
GeForce 50 Mobile
Process Size
5 nm
5 nm
Transistors
35,800 million
45,600 million
Die Size
294 mm²
378 mm²
Foundry
TSMC
TSMC
Density
121.8M / mm²
120.6M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.4
1.4
OpenCL
3.0
3.0
CUDA
8.9
12.0
Shader Model
6.8
6.9
Physical
Slot Width
Dual-slot
IGP
Length
240 mm 9.4 inches
Height
110 mm 4.3 inches
Outputs
1x HDMI 2.13x DisplayPort 1.4a
Portable Device Dependent
Bus Interface
PCIe 4.0 x16
PCIe 5.0 x16
Other
Launch Price
599 USD
Production
End-of-life
Active
Predecessor
GeForce 30
GeForce 40 Mobile
Successor
GeForce 50
View GeForce RTX 4070 Details View GeForce RTX 5090 Mobile Details