NVIDIA · Ada Lovelace
GeForce RTX 4070 SUPER vs NVIDIA GeForce RTX 3070 — Specs & Performance Compared
Side-by-side comparison of 2 GPU parts: GeForce RTX 4070 SUPER vs NVIDIA GeForce RTX 3070. All released in 2020. Specs, synthetic benchmarks, AI inference tok/s, and Amazon availability — click into any column header for the per-SKU benchmark page.
vs
NVIDIA · Ampere
NVIDIA GeForce RTX 3070
8 GB VRAM220W TDP$499 MSRP2020
Full NVIDIA GeForce RTX 3070 benchmarks →
*Price sourced from Amazon.com. Last updated 2026-05-27. Price and availability subject to change.
Specs side-by-side
| Spec | GeForce RTX 4070 SUPER | NVIDIA GeForce RTX 3070 |
|---|---|---|
| vram gb | — | 8 |
| vram type | — | GDDR6 |
| tdp w | — | 220 |
| cuda cores | — | 5888 |
Synthetic benchmarks
Higher is better. Per-row scores come from public benchmark databases (PassMark, TechPowerUp, Geekbench, Cinebench). Empty cells mean we don't yet have that benchmark on file for the SKU.
| Benchmark | GeForce RTX 4070 SUPER | NVIDIA GeForce RTX 3070 |
|---|---|---|
| 3DMark Fire Strike Extreme | — | 17,115 points |
| 3DMark Fire Strike Ultra | 12,166 points | — |
| 3DMark Port Royal | 13,011 points | 8,324 points |
| 3DMark Steel Nomad Graphics | — | 3,180 points |
| 3DMark Time Spy | — | 14,048 points |
| 3DMark Time Spy Extreme | 10,072 points | 6,907 points |
| PassMark G2D Mark | 1,193 pts | 1,005 pts |
| PassMark G3D Mark | 29,959 pts | 22,123 pts |
| Tom's Hardware GPU Hierarchy | 759 rank | — |
AI inference (tokens/sec)
Top community-reported tokens-per-second results for popular LLMs. See the per-SKU benchmark page for the full set including quantization, prompt length, and source citations.
GeForce RTX 4070 SUPER
- llama-2-7b: 108.7 tok/s
- qwen3:8b: 56.2 tok/s
- qwen3:14b: 37.1 tok/s
NVIDIA GeForce RTX 3070
- llama3:8b: 70.9 tok/s
- qwen3.5:9b: 57.9 tok/s
- qwen3:235b: 50.0 tok/s