Skip to main content
NVIDIA GeForce RTX 3070
NVIDIA · GPU · Ampere

NVIDIA GeForce RTX 3070 — Benchmarks & Specs

8 GB VRAM220W TDP$499 MSRP2020
Buy NVIDIA GeForce RTX 3070 on Amazon →

*Price sourced from Amazon.com. Price and availability subject to change.

Bottom line: how fast is the NVIDIA GeForce RTX 3070?

At 1440p (Ultra), the NVIDIA GeForce RTX 3070 averages 107 fps in Shadow of the Tomb Raider, per Tom's Hardware. For local LLM inference it generates 78.7 tokens/sec running llama2:7b at q4_0 under llama.cpp, per llama.cpp GitHub Discussion #10879. In PassMark G3D Mark it scores 22,123 pts, per PassMark. Its 8 GB of VRAM is the binding constraint for local inference: that capacity fits 7B-parameter models at Q4.

Every figure above is a row in the tables below, and each row links out to the review or public benchmark database the number was taken from. SpecPicks aggregates published measurements; it does not report first-party benchmark runs.

The NVIDIA GeForce RTX 3070 is a graphics card from the Ampere family released in 2020 from NVIDIA. Key on-paper specs include 8 GB of GDDR6 VRAM, 220W TDP. It launched with a $499 MSRP, though street prices typically diverge meaningfully from launch pricing — see the linked product cards below for current Amazon listings. Data on this page draws on 8+ Amazon listings, 10 synthetic benchmark results, 12 community AI inference reports, 14 measured game frame-rate results, aggregated from public benchmark databases (TechPowerUp, PassMark, Geekbench, Cinebench) and the LocalLLaMA community. Read this page when shopping the NVIDIA GeForce RTX 3070, comparing it against other graphics cards in your build, or sizing it for a specific workload (gaming at 1080p/1440p/4K, productivity benchmarks, or local LLM inference).

Gaming Performance (measured FPS)

Average and 1% low frame rates by game, resolution, and quality preset. Bars are scaled against the fastest result on this page.

Measured gaming frame rates for the NVIDIA GeForce RTX 3070 by game, resolution, and quality preset. “1% low” is the frame-time floor that determines perceived smoothness. Each row links to its original review or benchmark database.
Game Resolution Settings Relative Avg FPS 1% low Source
The Witcher 3: Wild Hunt (Next-Gen) 1080p Ultra 152 fps — AskGeek 2023-01-01
Cyberpunk 2077 1080p Ultra 107 fps — AskGeek 2023-10-01
Shadow of the Tomb Raider 1440p Ultra 107 fps — Tom's Hardware 2020-10-29
Forza Horizon 5 1080p Extreme 97 fps — Notebookcheck 2021-11-14
Final Fantasy XIV: Dawntrail 1440p Ultra 97 fps — Gamers Nexus 2025-04-17
Far Cry 6 1440p Ultra 94 fps 73 fps HardwareDB 2023-03-01
Cyberpunk 2077 1080p Ultra 85 fps — Gamers Nexus 2025-03-13
Forza Horizon 5 1440p Extreme 85 fps — Notebookcheck 2021-11-14
Dragon's Dogma 2 1080p Ultra 83 fps — Gamers Nexus 2025-04-17
Elden Ring 1440p Ultra 82 fps 60 fps HardwareDB 2023-03-01
Crimson Desert 1080p Ultra 81 fps — Gamers Nexus 2026-04-01
Baldur's Gate 3 1440p Ultra 80 fps — Gamers Nexus 2023-10-04
Assassin's Creed Valhalla 1440p Ultra 79 fps 54 fps HardwareDB 2023-03-01
A Plague Tale: Requiem 1440p Ultra 69 fps 52 fps HardwareDB 2023-03-01

AI Inference Performance

Tokens per second under each model + quantization. Higher = faster generation. Bars compare runs across the same model.

Local LLM inference throughput on the NVIDIA GeForce RTX 3070, in generated tokens per second. Higher is better; each row links to the community report or benchmark database it came from.
Model Quantization Relative Tokens/sec VRAM used Source
llama2:7b q4_0 llama.cpp 78.7 tok/s — llama.cpp GitHub Discussion #10879 2024-12-18
llama2:7b q4_0 llama.cpp 78.7 tok/s — llama.cpp GitHub Discussions 2024-07-01
llama3.1:8b q4_K_M ollama 73.7 tok/s 5.4 GB Patrick Hughes (bmdpat.com) 2026-07-26
llama3:8b q4_K_M llama.cpp 72.8 tok/s 4.6 GB GPU-Benchmarks-on-LLM-Inference (GitHub) 2024-07-01
llama3:8b q4_K_M llama.cpp 72.8 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub/XiongjieDai) 2024-05-01
llama3:8b q4_K_M llama.cpp 70.9 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub/XiongjieDai) 2024-05-01
llama3:8b q4_K_M llama.cpp 70.9 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub) 2024-05-01
llama3:8b q4_K_M llama.cpp 67.0 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub/XiongjieDai) 2024-05-01
llama3:8b q4_K_M llama.cpp 67.0 tok/s 4.6 GB GPU-Benchmarks-on-LLM-Inference (GitHub) 2024-07-01
llama3:8b q4_K_M llama.cpp 61.6 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub/XiongjieDai) 2024-05-01
llama3:8b q4_K_M llama.cpp 61.6 tok/s — GPU-Benchmarks-on-LLM-Inference (GitHub) 2024-07-01
llama3.1:8b q4_K_M llama.cpp 59.6 tok/s 8.0 GB LocalScore.ai 2024-12-01

Synthetic Benchmarks

Higher is better. Bars are scaled within each benchmark family (multi-thread, single-thread, etc.) so you can compare like-with-like at a glance.

Synthetic benchmark scores for the NVIDIA GeForce RTX 3070 — higher is better. Each row links to the public database the number was taken from.
Benchmark Relative Score Source
PassMark G3D Mark 22,123 pts PassMark 2026-04-20
3DMark Fire Strike Extreme 17,115 points KitGuru 2020-10-23
3DMark Fire Strike Extreme 17,115 points DigiStatement 2020-10-28
3DMark Time Spy 14,328 points 3DMark 2021-06-01
3DMark Time Spy 14,048 points KitGuru 2020-10-23
3DMark Time Spy 13,945 points DigiStatement 2020-10-28
3DMark Time Spy 13,945 points VideoCardz 2020-10-01
3DMark Fire Strike Ultra 8,749 points Guru3D 2020-10-23
3DMark Port Royal 8,412 points 3DMark 2022-04-11
3DMark Port Royal 8,324 points DigiStatement 2020-10-28

Products Featuring the NVIDIA GeForce RTX 3070

Full Specifications

tdp w220
vram gb8
vram typeGDDR6
cuda cores5888

NVIDIA GeForce RTX 3070 — Frequently Asked Questions

What is the NVIDIA GeForce RTX 3070 best used for?
NVIDIA GeForce RTX 3070 is positioned as a 8 GB VRAM Ampere-family graphics card. Use it for 1080p gaming and general productivity. See the synthetic + AI benchmark tables below for measured performance.
When was the NVIDIA GeForce RTX 3070 released, and what was its launch MSRP?
NVIDIA GeForce RTX 3070 launched in 2020 at a $499 MSRP. Street prices diverge from launch pricing over a product's lifetime — check the linked Amazon listings on this page for current availability.
Where do the benchmark numbers on this page come from?
Synthetic benchmarks are scraped from public databases (TechPowerUp, PassMark, Geekbench Browser, Cinebench leaderboards). AI inference numbers come from the LocalLLaMA community (Reddit threads, llama.cpp / Ollama discussion logs, and Phoronix when available). Every benchmark row carries an inline source citation — click through to verify the original number.
Can the NVIDIA GeForce RTX 3070 run local LLMs?
Yes — NVIDIA GeForce RTX 3070 has 12 AI inference benchmarks on file (see the AI Inference Performance section above for model + tokens-per-second numbers).
Where can I buy the NVIDIA GeForce RTX 3070?
Amazon listings are linked at the top of this page (in the hero CTA) and in the "Products Featuring this Hardware" section below. SpecPicks earns a small affiliate commission on qualifying purchases.

Buying guides that rank the NVIDIA GeForce RTX 3070's class

This page is the raw performance data. The guides below turn it into a ranked pick for a specific build.

Related GPUs

Editorial guides covering the NVIDIA GeForce RTX 3070

In-depth SpecPicks reviews, build guides, and head-to-heads referencing this graphics card.

More guides & deep dives from the SpecPicks archive

Browse all articles & guides →

More reviews from the SpecPicks archive

Browse all reviews →

More buying guides from SpecPicks

Browse all buying guides →
NVIDIA GeForce RTX 3070
NVIDIA GeForce RTX 3070
View on Amazon →