NVIDIA GeForce RTX 4070 AD103 vs NVIDIA GeForce RTX 4070 SUPER Comparison

Head-to-Head Benchmarks

The database records no direct head-to-head benchmark results for the NVIDIA GeForce RTX 4070 AD103 against the NVIDIA GeForce RTX 4070 SUPER. The RTX 4070 AD103 has an empty benchmark array, meaning no scores are available for any workload. The RTX 4070 SUPER, by contrast, has a full set of recorded measurements across multiple APIs and workloads.

The RTX 4070 SUPER delivers an average benchmark score of 43,223 across its tested suite. This places it at the 83rd percentile among all GPUs tracked in the database. The RTX 4070 AD103 has a percentile rank of 50, but with no recorded scores, this rank reflects the absence of data rather than measured performance. The delta between the two cards cannot be computed from direct comparison, as no matching test results exist.

Looking at the RTX 4070 SUPER's individual scores, the strongest result appears in Geekbench Vulkan with 205,624 points. The OpenCL result follows at 172,795. Passmark G3D records 29,995, while Passmark GPU Compute shows 17,108. The smaller Passmark scores include DirectX 9 at 344, DirectX 11 at 273, DirectX 10 at 167, and DirectX 12 at 110. The 2D graphics score sits at 1,184. The 3DMark Steel Nomad DX12 test produces 4,627.

The RTX 4070 SUPER's nearest rivals in the database provide context for its standing. The NVIDIA Quadro M6000 24 GB scores 43,262, which is 0.1% lower. The NVIDIA GeForce RTX 5050 Mobile scores 43,268, also 0.1% lower. The NVIDIA Quadro M6000 scores 43,301, 0.2% lower. The NVIDIA GeForce RTX 4090 Mobile scores 43,667, which is 1% higher. These tight margins place the RTX 4070 SUPER in a narrow band of comparable performance.

Since the RTX 4070 AD103 has no benchmark entries, any numerical comparison must rely on architectural and specification differences rather than measured scores. The RTX 4070 SUPER's recorded performance stands as the only quantified data point for either card in this comparison.

Where Each One Wins

Without recorded benchmarks for the RTX 4070 AD103, the use-case split rests on the available scores for the RTX 4070 SUPER and the specification differences between the two cards.

The RTX 4070 SUPER shows strength in compute-heavy workloads. Its Geekbench OpenCL score of 172,795 and Vulkan score of 205,624 indicate solid cross-platform compute capability. The Passmark GPU Compute result of 17,108 further supports this pattern. The 3DMark Steel Nomad DX12 score of 4,627 confirms the card handles modern DirectX 12 rendering tasks effectively.

The RTX 4070 SUPER also performs well in legacy DirectX workloads, with Passmark DirectX 9 scoring 344 and DirectX 11 scoring 273. The lower DirectX 12 score of 110 suggests that the newer API path may not be as heavily optimized in that particular Passmark test, but the 3DMark result counters this with a strong DX12 showing.

For the RTX 4070 AD103, the absence of scores means no workload can be identified as a win. The card's specifications, however, reveal a different configuration. It uses the AD103 chip with 5,888 shading units, 184 texture mapping units, and 64 raster operation units. The RTX 4070 SUPER uses the AD104 chip with 7,168 shading units, 224 TMUs, and 80 ROPs. The SUPER has more of every execution resource, which suggests it would outperform in most rasterization and compute tasks, but the database lacks the measurements to confirm this.

The RTX 4070 SUPER's pixel rate of 198.0 GPixel/s and texture rate of 554.4 GTexel/s exceed the RTX 4070 AD103's 158.4 GPixel/s and 455.4 GTexel/s. These rates point to advantages in fill-rate-bound and texture-heavy scenes, but again, no direct benchmark confirms the real-world margin.

Architecture Differences

Both cards use the Ada Lovelace architecture and are fabricated on TSMC's 5 nm process. The RTX 4070 AD103 carries the AD103 chip, while the RTX 4070 SUPER uses the AD104 chip. This is the primary architectural distinction between the two.

The RTX 4070 AD103 packs 45,900 million transistors on a 379 mm² die. The RTX 4070 SUPER has 35,800 million transistors on a 294 mm² die. Despite the AD103's larger transistor count and die size, the SUPER's AD104 achieves a slightly higher transistor density at 121.8 million per mm² versus 121.1 million per mm². The AD103 chip is physically larger, but the AD104 is more densely packed.

Both cards feature 46 ray tracing cores on the AD103 versus 56 on the AD104? No, the data shows the RTX 4070 AD103 has 46 RT cores and 184 tensor cores. The RTX 4070 SUPER has 56 RT cores and 224 tensor cores. The SUPER's higher counts in both categories indicate more dedicated hardware for ray tracing and AI-accelerated workloads.

The memory subsystem is identical between the two cards. Both use 12 GB of GDDR6X memory on a 192-bit bus, with a memory clock of 1313 MHz and 21 Gbps effective speed. Both deliver 504.2 GB/s of bandwidth. The memory architecture does not differentiate these cards.

The API support matches exactly. Both cards expose DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4. Neither card has a distinct API advantage.

The process node is the same 5 nm TSMC node for both. The foundry is identical. The architectural difference comes down to chip selection and the resulting resource counts, not fabrication technology.

Specification Differences

The two cards differ in several specification fields. The chip is different: AD103 for the RTX 4070 AD103, AD104 for the RTX 4070 SUPER. The transistor count differs, with the AD103 at 45,900 million and the AD104 at 35,800 million. Die size differs at 379 mm² versus 294 mm². Transistor density is nearly identical at 121.1M per mm² and 121.8M per mm².

Base clock speeds differ. The RTX 4070 AD103 runs at 1920 MHz base, while the RTX 4070 SUPER runs at 1980 MHz base. Boost clocks are identical at 2475 MHz for both. Memory clocks match at 1313 MHz with 21 Gbps effective.

Shading units differ significantly: 5,888 on the AD103 versus 7,168 on the SUPER. TMUs differ at 184 versus 224. ROPs differ at 64 versus 80. RT cores differ at 46 versus 56. Tensor cores differ at 184 versus 224.

Pixel rate differs at 158.4 GPixel/s for the AD103 and 198.0 GPixel/s for the SUPER. Texture rate differs at 455.4 GTexel/s versus 554.4 GTexel/s. FP32 compute differs at 29.15 TFLOPS versus 35.48 TFLOPS. FP16 matches FP32 at a 1:1 ratio for both.

TDP differs at 200 W for the AD103 and 220 W for the SUPER. Both are dual-slot cards with a single 16-pin power connector. Both recommend a 550 W power supply. Both use PCIe 4.0 x16. Display outputs are identical: 1x HDMI 2.1 and 3x DisplayPort 1.4a.

Dimensions differ. The RTX 4070 AD103 measures 240 mm in length, 110 mm in height, and 40 mm in width. The RTX 4070 SUPER measures 267 mm in length, 112 mm in height, and 42 mm in width. The SUPER is longer, taller, and thicker.

Release dates differ. The RTX 4070 AD103 launched on 2024-02-29. The RTX 4070 SUPER launched on 2024-01-16. Both are end-of-life products. Both have a launch MSRP of 599 USD. Both belong to the GeForce 40 series, with the GeForce 30 as predecessor and GeForce 50 as successor.

FAQ

Q: Which card has more shading units?

A: The NVIDIA GeForce RTX 4070 SUPER has 7,168 shading units, compared to 5,888 on the NVIDIA GeForce RTX 4070 AD103.

Q: Do the two cards use the same memory configuration?

A: Yes, both cards use 12 GB of GDDR6X memory on a 192-bit bus with 504.2 GB/s bandwidth and a 1313 MHz memory clock.

Q: What is the average benchmark score for the RTX 4070 SUPER?

A: The RTX 4070 SUPER records an average benchmark score of 43,223 and sits at the 83rd percentile among all GPUs in the database.

Q: Is there any benchmark data for the RTX 4070 AD103?

A: No, the database contains no benchmark scores for the RTX 4070 AD103. Its average benchmark score is recorded as 0 and its percentile rank is 50.

Q: Which card has the higher FP32 compute throughput?

A: The RTX 4070 SUPER delivers 35.48 TFLOPS of FP32 compute, compared to 29.15 TFLOPS for the RTX 4070 AD103.

Q: How close is the RTX 4070 SUPER to its nearest rival in the database?

A: The closest rival is the NVIDIA Quadro M6000 24 GB with an average score of 43,262, which is 0.1% lower than the RTX 4070 SUPER's 43,223.

The Verdict

The recorded data shows a clear split. The RTX 4070 SUPER has a full set of benchmark measurements and a strong percentile rank of 83. The RTX 4070 AD103 has no benchmark data at all, making any direct performance verdict impossible from measurements alone.

For buyers selecting based on recorded performance, the RTX 4070 SUPER is the only card with quantified results. Its average score of 43,223 places it among capable performers, with nearest rivals within 1% in either direction. The RTX 4070 SUPER also carries higher execution resources across every category: more shading units, more TMUs, more ROPs, more RT cores, more tensor cores, and higher pixel, texture, and compute rates.

The RTX 4070 AD103 offers a larger die at 379 mm² and more transistors at 45,900 million, but it does not translate that into higher resource counts. The SUPER's AD104 die is smaller at 294 mm² and uses fewer transistors at 35,800 million, yet delivers more of every functional unit. The SUPER also boosts to the same 2475 MHz while starting from a higher base clock of 1980 MHz versus 1920 MHz.

The RTX 4070 SUPER consumes 220 W versus 200 W for the AD103, and it is physically larger in all three dimensions. Both cards share the same memory, bandwidth, API support, power connector, PSU recommendation, and launch MSRP of 599 USD.

For anyone choosing strictly from the data, the RTX 4070 SUPER is the measured performer with the higher resource counts and the only recorded benchmark scores. The RTX 4070 AD103 lacks any quantified evidence of performance in the database, so it cannot be recommended on the basis of measured results. The SUPER is the card with the data to back its position.