GeForce RTX 3080 vs GeForce RTX 4070: specs and benchmark comparison
BNVIDIA · Ada Lovelace
NVIDIA GeForce RTX 4070
- 12 GB GDDR6 / GDDR6X
- 200 W
- PSU 650 W+
April 13, 2023 · $599 at launch
Short verdict
NVIDIA GeForce RTX 4070 leads in every scenario with comparable benchmarks: rendering and compute. This does not imply an advantage in tasks without benchmarks.
Compare with AIRendering and compute
B +27.6%NVIDIA GeForce RTX 4070 leads in this scenario (up to 27.6% in comparable benchmarks).
No comparable benchmarks yet: gaming: rasterization, gaming: ray tracing, power efficiency.
Advantages
Only measured and rated values: a noticeable benchmark gap and specs where more is better.
Where GeForce RTX 3080 is better1
Memory bus 128 bit more320 bit vs 192 bit
Where GeForce RTX 4070 is better2
Blender Benchmark 4.5.0 15.4% faster5,226 samples/min vs 4,527 samples/min
Video memory 2 GB more12 GB vs 10 GB
Benchmarks
Results are compared within one benchmark, one version and the same conditions.
- B ahead by 15.4%
Blender Benchmark 4.5.0
GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) · samples/min, higher is better
GeForce RTX 30804,526.6 samples/minGeForce RTX 40705,225.8 samples/minPosition among tested models (67, OptiX)
A · 4,527B · 5,22605,00010,00015,000Each dot is one model’s median in this benchmark. The scale starts at zero, samples/min. NVIDIA GeForce RTX 4070 is ahead by 15.4% in this benchmark.
About the measurement and source
- A · NVIDIA GeForce RTX 3080
4,526.6 samples/min · community measurement, sample 536 · November 14, 2026
Blender Open Data (Blender Foundation)
Median of 536 submissions (43 outliers excluded), interquartile range 4378.6–4702.09; submissions 2025-07-17 to 2026-09-13; OPTIX backend. Snapshot 2026-11-14.
- B · NVIDIA GeForce RTX 4070
5,225.8 samples/min · community measurement, sample 248 · November 14, 2026
Blender Open Data (Blender Foundation)
Median of 248 submissions (14 outliers excluded), interquartile range 5058.13–5339.47; submissions 2025-07-17 to 2026-07-07; OPTIX backend. Snapshot 2026-11-14.
difference = |A − B| ÷ the smaller value × 100%. The 5% noticeable threshold is a UX heuristic, not statistical significance.
Cycles render in Blender Benchmark. Median of the sum of samples per minute over three scenes after removing outliers.
Other comparable benchmarks (12)
| Benchmark | A | B | Difference |
|---|---|---|---|
| Blender Benchmark 4.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,450.1 | 5,153.6 | B +15.8% |
| Blender Benchmark 5.1.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,379.9 | 5,265.7 | B +20.2% |
| Blender Benchmark 5.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,197.4 | 5,114.2 | B +21.8% |
| Blender Benchmark 3.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,025.1 | 5,812 | B +15.7% |
| Blender Benchmark 3.4.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,001.5 | 5,890.6 | B +17.8% |
| Blender Benchmark 3.5.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,251.5 | 5,999.6 | B +14.2% |
| Blender Benchmark 3.6.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,201.5 | 6,034.8 | B +16% |
| Blender Benchmark 4.0.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,564.4 | 5,403.1 | B +18.4% |
| Blender Benchmark 4.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,614.7 | 5,432 | B +17.7% |
| Blender Benchmark 4.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,511.6 | 5,113.6 | B +13.3% |
| Blender Benchmark 4.4.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,478.6 | 5,172.3 | B +15.5% |
| Blender Benchmark 5.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,330.8 | 5,526.5 | B +27.6% |
Each benchmark version has its own scale: numbers are compared only within a row.
Reference results (70)
| Benchmark | A | B | Difference |
|---|---|---|---|
| Blender Benchmark 2.90GPU, bmw27 (CUDA) | 25.5 | — | no pair |
| Blender Benchmark 2.90.1GPU, bmw27 (CUDA) | 25.4 | — | no pair |
| Blender Benchmark 2.91.2GPU, bmw27 (CUDA) | 21.7 | — | no pair |
| Blender Benchmark 2.92GPU, bmw27 (CUDA) | 21.8 | — | no pair |
| Blender Benchmark 2.93.1GPU, bmw27 (CUDA) | 19.9 | — | no pair |
| Blender Benchmark 2.90GPU, bmw27 (OPTIX) | 10.7 | — | no pair |
| Blender Benchmark 2.91.2GPU, bmw27 (OPTIX) | 11 | — | no pair |
| Blender Benchmark 2.92GPU, bmw27 (OPTIX) | 10.7 | — | no pair |
| Blender Benchmark 2.93.1GPU, bmw27 (OPTIX) | 11.4 | — | no pair |
| Blender Benchmark 2.90GPU, classroom (CUDA) | 78.1 | — | no pair |
| Blender Benchmark 2.90.1GPU, classroom (CUDA) | 96.6 | — | no pair |
| Blender Benchmark 2.91.2GPU, classroom (CUDA) | 68.6 | — | no pair |
| Blender Benchmark 2.92GPU, classroom (CUDA) | 65.3 | — | no pair |
| Blender Benchmark 2.93.1GPU, classroom (CUDA) | 93.8 | — | no pair |
| Blender Benchmark 2.90GPU, classroom (OPTIX) | 41.4 | — | no pair |
| Blender Benchmark 2.91.2GPU, classroom (OPTIX) | 42.5 | — | no pair |
| Blender Benchmark 2.92GPU, classroom (OPTIX) | 39.5 | — | no pair |
| Blender Benchmark 2.93.1GPU, classroom (OPTIX) | 39.9 | — | no pair |
| Blender Benchmark 2.90GPU, fishy_cat (CUDA) | 45.8 | — | no pair |
| Blender Benchmark 2.90.1GPU, fishy_cat (CUDA) | 46.4 | — | no pair |
| Blender Benchmark 2.91.2GPU, fishy_cat (CUDA) | 39.5 | — | no pair |
| Blender Benchmark 2.92GPU, fishy_cat (CUDA) | 40.8 | — | no pair |
| Blender Benchmark 2.93.1GPU, fishy_cat (CUDA) | 40.2 | — | no pair |
| Blender Benchmark 2.90GPU, fishy_cat (OPTIX) | 19.8 | — | no pair |
| Blender Benchmark 2.91.2GPU, fishy_cat (OPTIX) | 19.7 | — | no pair |
| Blender Benchmark 2.92GPU, fishy_cat (OPTIX) | 21.4 | — | no pair |
| Blender Benchmark 2.93.1GPU, fishy_cat (OPTIX) | 21.6 | — | no pair |
| Blender Benchmark 2.90GPU, koro (CUDA) | 68.4 | — | no pair |
| Blender Benchmark 2.90.1GPU, koro (CUDA) | 68.4 | — | no pair |
| Blender Benchmark 2.91.2GPU, koro (CUDA) | 68.4 | — | no pair |
| Blender Benchmark 2.92GPU, koro (CUDA) | 68.9 | — | no pair |
| Blender Benchmark 2.93.1GPU, koro (CUDA) | 64.6 | — | no pair |
| Blender Benchmark 2.90GPU, koro (OPTIX) | 42.7 | — | no pair |
| Blender Benchmark 2.91.2GPU, koro (OPTIX) | 42.1 | — | no pair |
| Blender Benchmark 2.92GPU, koro (OPTIX) | 45.8 | — | no pair |
| Blender Benchmark 2.93.1GPU, koro (OPTIX) | 45.3 | — | no pair |
| Blender Benchmark 2.90GPU, pavillon_barcelona (CUDA) | 145.3 | — | no pair |
| Blender Benchmark 2.90.1GPU, pavillon_barcelona (CUDA) | 139.1 | — | no pair |
| Blender Benchmark 2.91.2GPU, pavillon_barcelona (CUDA) | 129.7 | — | no pair |
| Blender Benchmark 2.92GPU, pavillon_barcelona (CUDA) | 130 | — | no pair |
| Blender Benchmark 2.93.1GPU, pavillon_barcelona (CUDA) | 119 | — | no pair |
| Blender Benchmark 2.90GPU, pavillon_barcelona (OPTIX) | 53.6 | — | no pair |
| Blender Benchmark 2.91.2GPU, pavillon_barcelona (OPTIX) | 54.5 | — | no pair |
| Blender Benchmark 2.92GPU, pavillon_barcelona (OPTIX) | 53 | — | no pair |
| Blender Benchmark 2.93.1GPU, pavillon_barcelona (OPTIX) | 54.2 | — | no pair |
| Blender Benchmark 2.90GPU, victor (CUDA) | 191.6 | — | no pair |
| Blender Benchmark 2.90.1GPU, victor (CUDA) | 195 | — | no pair |
| Blender Benchmark 2.91.2GPU, victor (CUDA) | 150.7 | — | no pair |
| Blender Benchmark 2.92GPU, victor (CUDA) | 207.5 | — | no pair |
| Blender Benchmark 2.93.1GPU, victor (CUDA) | 165.8 | — | no pair |
| Blender Benchmark 2.90GPU, victor (OPTIX) | 60.7 | — | no pair |
| Blender Benchmark 2.91.2GPU, victor (OPTIX) | 62.7 | — | no pair |
| Blender Benchmark 2.92GPU, victor (OPTIX) | 70.9 | — | no pair |
| Blender Benchmark 2.93.1GPU, victor (OPTIX) | 71.7 | — | no pair |
| Blender Benchmark 3.1.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 2,981.3 | — | no pair |
| Blender Benchmark 3.2.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 2,845.2 | — | no pair |
| Blender Benchmark 3.3.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 2,902.8 | — | no pair |
| Blender Benchmark 3.5.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 3,291.3 | — | no pair |
| Blender Benchmark 3.6.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 3,193.9 | 3,719.9 | reference |
| Blender Benchmark 4.0.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | 2,989.8 | 3,435.7 | reference |
| Blender Benchmark 3.0.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,065.1 | — | no pair |
| Blender Benchmark 3.1.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,893.1 | — | no pair |
| Blender Benchmark 3.1.2GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 5,100.4 | — | no pair |
| Blender Benchmark 3.2.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,851.8 | — | no pair |
| Blender Benchmark 3.2.1GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,870.7 | 5,138.5 | reference |
| Blender Benchmark 5.0.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,105.5 | 5,108.7 | reference |
| Blender Benchmark 5.3.0GPU, sum of scenes monster + junkshop + classroom (OptiX / HIP / oneAPI / Metal) | 4,546.6 | 5,906.2 | reference |
| Blender Benchmark 4.2.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | — | 3,387.9 | no pair |
| Blender Benchmark 4.5.0GPU, sum of scenes monster + junkshop + classroom (CUDA) | — | 3,291.8 | no pair |
| Blender Benchmark 5.1.1GPU, sum of scenes monster + junkshop + classroom (CUDA) | — | 3,114.6 | no pair |
Each benchmark version has its own scale: numbers are compared only within a row.
Specifications
| Specification | A · GeForce RTX 3080 | B · GeForce RTX 4070 | Difference |
|---|---|---|---|
| Chip | |||
| Architecture | Ampere | Ada Lovelace | differs |
| Shader unitsShader unit counts are comparable only within one architecture. | 8,704 (CUDA cores) | 5,888 (CUDA cores) | not directly comparable |
| Clocks | |||
| Base clock | 1,440 MHz | 1,920 MHz | B: 480 MHz more |
| Boost clockClocks of different architectures do not reflect performance directly. For laptop chips, the clock depends on the laptop’s power limit. | 1,710 MHz | 2,480 MHz | B: 770 MHz more |
| Memory | |||
| Video memory | 10 GB | 12 GB | B: 2 GB more |
| Memory type | GDDR6X | GDDR6 / GDDR6X | differs |
| Memory bus | 320 bit | 192 bit | A: 128 bit more |
| Power and dimensions | |||
| Board powerTGP (NVIDIA) and TBP (AMD/Intel) are calculated differently; higher is not better. For laptops, the chip maker sets the range and the laptop maker picks the value, so the same chip behaves differently in different laptops. | 320 W (board power) | 200 W (TGP) | not directly comparable |
| Recommended PSU | 750 W | 650 W | A: 100 W higher |
| Interface | PCIe 4.0 | no data | no data for B |
| Length (reference / Founders Edition)Partner cards have different dimensions; check the specific SKU. | 285 mm | 244 mm | A: 41 mm higher |
| Thickness (slots) | 2 | 2 | equal |
| Power connectors (reference) | 2x 8-pin (adapter to 1x 12-pin included) — Founders Edition | no data | no data for B |
Chip
Architecture
AAmpere
BAda Lovelace
differsShader units
A8,704 (CUDA cores)
B5,888 (CUDA cores)
not directly comparableShader unit counts are comparable only within one architecture.
Clocks
Base clock
A1,440 MHz
B1,920 MHz
B: 480 MHz moreBoost clock
A1,710 MHz
B2,480 MHz
B: 770 MHz moreClocks of different architectures do not reflect performance directly. For laptop chips, the clock depends on the laptop’s power limit.
Memory
Video memory
A10 GB
B12 GB
B: 2 GB moreMemory type
AGDDR6X
BGDDR6 / GDDR6X
differsMemory bus
A320 bit
B192 bit
A: 128 bit more
Power and dimensions
Board power
A320 W (board power)
B200 W (TGP)
not directly comparableTGP (NVIDIA) and TBP (AMD/Intel) are calculated differently; higher is not better. For laptops, the chip maker sets the range and the laptop maker picks the value, so the same chip behaves differently in different laptops.
Recommended PSU
A750 W
B650 W
A: 100 W higherInterface
APCIe 4.0
Bno data
no data for BLength (reference / Founders Edition)
A285 mm
B244 mm
A: 41 mm higherPartner cards have different dimensions; check the specific SKU.
Thickness (slots)
A2
B2
equalPower connectors (reference)
A2x 8-pin (adapter to 1x 12-pin included) — Founders Edition
Bno data
no data for B
“More” does not always mean “better”: higher power draw or size is a system requirement, not an advantage (such bars are dimmed).
Compare with AI
An AI model reviews the specs and benchmarks of this pair and gives a plain-language verdict on who each model suits.
Questions and answers
How much video memory do NVIDIA GeForce RTX 3080 and NVIDIA GeForce RTX 4070 have?
NVIDIA GeForce RTX 3080: 10 GB GDDR6X, 320-bit bus. NVIDIA GeForce RTX 4070: 12 GB GDDR6 / GDDR6X, 192-bit bus.
What power supply does the manufacturer recommend?
For NVIDIA GeForce RTX 3080, 750 W or more; for NVIDIA GeForce RTX 4070, 650 W or more. This is a recommendation for a typical system; also check the power connectors of the specific card model.
Which is faster: NVIDIA GeForce RTX 3080 or NVIDIA GeForce RTX 4070?
NVIDIA GeForce RTX 4070 leads in every scenario with comparable benchmarks: rendering and compute. This does not imply an advantage in tasks without benchmarks.
See also
Data sources (3)
- NVIDIA — specifications of NVIDIA GeForce RTX 3080, specifications of NVIDIA GeForce RTX 4070checked November 7, 2026
- Blender Open Data (Blender Foundation) — benchmark Blender Benchmark 2.90, benchmark Blender Benchmark 2.90.1, benchmark Blender Benchmark 2.91.2, benchmark Blender Benchmark 2.92, benchmark Blender Benchmark 2.93.1, benchmark Blender Benchmark 3.6.0, benchmark Blender Benchmark 4.0.0, benchmark Blender Benchmark 4.2.0, benchmark Blender Benchmark 4.5.0, benchmark Blender Benchmark 5.1.1, benchmark Blender Benchmark 3.1.0, benchmark Blender Benchmark 3.2.0, benchmark Blender Benchmark 3.3.0, benchmark Blender Benchmark 3.5.0, benchmark Blender Benchmark 5.2.0, benchmark Blender Benchmark 3.2.1, benchmark Blender Benchmark 3.4.0, benchmark Blender Benchmark 4.1.0, benchmark Blender Benchmark 4.3.0, benchmark Blender Benchmark 4.4.0, benchmark Blender Benchmark 5.0.0, benchmark Blender Benchmark 5.1.0, benchmark Blender Benchmark 5.3.0, benchmark Blender Benchmark 3.0.1, benchmark Blender Benchmark 3.1.2checked November 14, 2026
- Wikipedia (English) — release date and launch pricechecked November 14, 2026
How we compare: methodology. Where the data comes from: sources.