Tensor G3 vs A14 Bionic
We compared the 9-core Google Tensor G3 (Mali-G715 MP7) with the older 6-core Apple A14 Bionic (Apple A14 GPU) SoC. Here you will find the pros and cons of each chip, technical specs, and comprehensive tests in benchmarks, like AnTuTu and Geekbench.
Review
CPU Performance
Single and multi-core processor tests
Gaming Performance
GPU performance in games and OpenCL/Vulkan
Battery life
Efficiency of battery consumption
NanoReview Score
Overall chip score
Key Differences
Pros of Google Tensor G3
- Performs 2.4x better in floating-point computations
- Supports 100% higher memory bandwidth (68.2 against 34.1 GB/s)
- Has 3 more cores
- Has a smaller size transistor (4 versus 5 nm)
- Better instruction set architecture
- Shows better (up to 6%) AnTuTu 11 score β 1409K vs 1326K
Pros of Apple A14 Bionic
- Higher GPU frequency (~44%)
- 7% higher CPU clock speed (3100 vs 2910 MHz)
Benchmarks
Performance tests in popular benchmarksAnTuTu 11
| CPU | 482403 | 364692 |
| GPU | 384611 | 311592 |
| Memory | 218603 | 180518 |
| UX | 323399 | 470169 |
| Total score | 1409016 | 1326975 |
| Samples | [33 tests] | [55 tests] |
AnTuTu results from iOS and Android are not directly comparable! They won't affect NanoReview scores.
GeekBench 6
Single-Core Score
Multi-Core Score
| Asset compression | 197.6 MB/sec | 182.6 MB/sec |
| HTML 5 Browser | 134.7 pages/sec | 110.6 pages/sec |
| PDF Renderer | 148.4 Mpixels/sec | 123.2 Mpixels/sec |
| Image detection | 82.1 images/sec | 104.9 images/sec |
| HDR | 142.1 Mpixels/sec | 164.1 Mpixels/sec |
| Background blur | 13.2 images/sec | 19.1 images/sec |
| Photo processing | 48.4 images/sec | 50.8 images/sec |
| Ray tracing | 5.34 Mpixels/sec | 5.36 Mpixels/sec |
| Samples | 11 tests | 22 tests |
Compute Score (GPU)
| Background blur | 10.6 images/sec | 27.6 images/sec |
| Face detection | 6.85 images/sec | 35.2 images/sec |
| Horizon detection | 0.16 Gpixels/sec | 0.54 Gpixels/sec |
| Edge detection | 0.35 Gpixels/sec | 0.79 Gpixels/sec |
| Gaussian blur | 0.32 Gpixels/sec | 0.83 Gpixels/sec |
| Feature matching | 0.08 Gpixels/sec | 0.24 Gpixels/sec |
| Stereo matching | 11.9 Gpixels/sec | 50.5 Gpixels/sec |
| Particle physics | 1088 FPS | 1772.1 FPS |
| API | Vulkan | Metal |
| Samples | 9 tests | 17 tests |
Compute results use different APIs (Metal vs Vulkan/OpenCL) and are not directly comparable!
3DMark
3DMark Steel Nomad Light
| Score | 982 | - |
| Stability | 74% | - |
| Graphics test | 7 FPS | - |
3DMark Solar Bay
| Score | - | 2767 |
| Stability | - | 85% |
| Graphics test | - | 11 FPS |
| Section 1 | - | 13 FPS |
| Section 2 | - | 10 FPS |
| Section 3 | - | 8 FPS |
3DMark Wild Life Extreme
| Score | - | 2229 |
| Stability | - | 86% |
| Graphics test | - | 13 FPS |
Recent User Tests
Google Tensor G3
| Date | Benchmark | Result |
|---|---|---|
| π 2026-08-28 (asfiii-malik) | AnTuTu 11 | 1515235 |
| π 2026-08-28 (asfiii-baloch) | AnTuTu 11 | 1515235 |
| π 2026-05-05 (s1mpick) | AnTuTu 11 | 1338293 |
| π 2026-05-04 (reazz) | AnTuTu 11 | 1356451 |
| π 2026-05-04 (Bilal Ahmed) | AnTuTu 11 | 1367231 |
Apple A14 Bionic
| Date | Benchmark | Result |
|---|---|---|
| π 2026-09-26 (kudryash) | AnTuTu 11 | 1196958 |
| π 2026-08-04 (Stelle) | AnTuTu 11 | 1288870 |
| π 2026-06-30 (Kirious) | AnTuTu 11 | 1251875 |
| π 2026-06-30 (Kirious) | AnTuTu 11 | 1251875 |
| π 2026-06-29 (Kirillkryt) | AnTuTu 11 | 1284629 |
Specifications
Full list of technical specifications of Tensor G3 and A14 BionicCPU
| Architecture | 1x 2.91 GHz β Cortex-X3 4x 2.37 GHz β Cortex-A715 4x 1.7 GHz β Cortex-A510 |
2x 3.1 GHz β Firestorm 4x 1.8 GHz β Icestorm |
| Cores | 9 | 6 |
| Frequency | 2910 MHz | 3100 MHz |
| Instruction set | ARMv9-A | ARMv8.5-A |
| L1 cache | - | 128 KB |
| L2 cache | - | 8 MB |
| Process | 4 nanometers | 5 nanometers |
| Transistor count | - | 11.8 billion |
| TDP (Sustained Power Limit) | 6 W | 6 W |
| Manufacturing | Samsung | TSMC |
Graphics
| GPU name | Mali-G715 MP7 | Apple A14 GPU |
| Architecture | Valhall 4th gen | Apple Bionic GPU |
| GPU frequency | 890 MHz | 1278 MHz |
| Pipelines | 7 | 4 |
| Shading units | 128 | 64 |
| Total shaders | 896 | 256 |
| FLOPS | 1594.8 Gigaflops | 654.3 Gigaflops |
| Vulkan version | 1.3 | - |
| OpenCL version | 2.0 | - |
AI Accelerator
| Neural processor (NPU) | Yes | Apple Neural Engine |
| NPU theoretical performance | - | 11 TOPS |
Memory
| Memory type | LPDDR5X | LPDDR4X |
| Memory frequency | 4266 MHz | 2133 MHz |
| Bus | 4x 16 Bit | 4x 16 Bit |
| Max bandwidth | 68.2 Gb/s | 34.1 Gb/s |
| Max size | 16 GB | 6 GB |
Multimedia (ISP)
| Storage type | UFS 3.1, UFS 4.0 | NVMe |
| Max display resolution | 3840 x 2160 | 2732 x 2048 |
| Max camera resolution | 1x 200MP, 2x 32MP | - |
| Video capture | 8K at 30FPS, 4K at 120FPS | 4K at 60FPS |
| Video playback | 8K at 30FPS, 4K at 120FPS | 4K at 60FPS |
| Video codecs | - H.264 - H.265 - AV1 - VP9 |
- H.264 - H.265 - Motion JPEG |
| Audio codecs | - AAC - AIFF - CAF - MP3 - MP4 - WAV |
- AAC - AAX - AAX+ - AC-3 - AIFF - CAF - E-AC-3 - MP3 - MP4 - WAV |
Connectivity
| Modem | Exynos 5300i | - |
| 4G support | LTE Cat. 24 | LTE Cat. 18 |
| 5G support | Yes | Yes |
| Download speed | Up to 7300 Mbps | Up to 7500 Mbps |
| Upload speed | Up to 3700 Mbps | Up to 2500 Mbps |
| Wi-Fi | 7 | 6 |
| Bluetooth | 5.3 | 5.0 |
| Navigation | GPS, GLONASS, Beidou, Galileo, QZSS | GPS, GLONASS, Beidou, Galileo |
Info
| Announced | October 2023 | September 2020 |
| Class | Flagship | Flagship |
| Model number | GS301 (S5P9865) | APL1W01 |
| Official page | Google Tensor G3 official site | - |
| Successor | - Google Tensor G5 | - Apple A19 Pro |
Cast your vote
Total votes: 178