buyaigear
Prices checked 7 min ago Sign in

NVIDIA · Ada Lovelace · 12 GB · Qwen3.8 27B · Q4_K_M at 8K context: ~5.2 tok/s with RAM offload

NVIDIA GeForce RTX 4070 12GB

The efficient 12 GB Ada card: 200 W, FP8 tensor cores, 504 GB/s. An 8B model at Q8 or a 14B at Q4 fits; slower than a 3080 Ti but it sips power and is still sold new.

Watch Qwen3.8 27B stream at this machine's speed

Lowest price right now

$569.99

eBay · free shipping · $47 per GB · observed 2026-11-14 18:37 UTC

Buy the cheapest on eBay, $569.99

free shipping · New · newegg 1,087,106 · 99.6%

See more on eBay

Or rent one from $0.10 an hour

Price alert

One email per move, item plus shipping. Clear the price to hear about any drop. Manage alerts.

Price history

Daily low across every tracked vendor, last 6 days.

$516$529$541$554Nov 9Nov 14lowest $518.42

Where to buy

VendorPriceConditionObserved
eBay
cheapest active listing		 $569.99
free shipping		 new and used 2026-11-14 18:37 UTC Go to eBay

On eBay now

21 listings passed our filters at the last check, 7 min ago. We check every 28 minutes. Cheapest delivered first. Seller reputation is feedback count and positive rate.

All 21 listings · Open the search on eBay

Rent one by the hour

Try it before you buy, or skip buying: the cheapest on-demand price for one RTX 4070 in the cloud right now, checked every hour.

CloudPer hourListed asAvailable
Vast.ai$0.10RTX 40707 verified offersRent on Vast.ai

At $0.10 an hour on Vast.ai, the $569.99 a used one costs on eBay buys about 6,000 hours of rental: 8 months around the clock, or 2.1 years at 8 hours a day. Use it more than that and buying comes out ahead; your power bill pushes that point later, resale value pulls it earlier.

Rental prices for every card · Rental links are referral links: we earn a small share of what you spend, at no cost to you.

What it runs

For each model, grouped by vendor with the newest first: the largest quant that fits with 8K of context and the estimated speed for one user, at 8K and at 32K of context. On a card, a model too big for its memory can still run with part of the weights in system RAM; we assume 64 GB, and the speed shown accounts for it. The model you picked on the ladder is highlighted. Estimates from the spec sheet, not measurements: see how we estimate.

ModelQuantNeeds at 8KFitMax contextDecode at 8KAt 32KPrefill tok/s
DeepSeek
DeepSeek V4.1 Flash 552B-A16B Sep 2026Q4_K_M338 GBdoes not fit
DeepSeek V4 Flash 284B-A13B Jul 2026Q4_K_M175 GBdoes not fit
DeepSeek V3.1 671B-A37B Aug 2025Q4_K_M411 GBdoes not fit
Gemma
Gemma 4 12B May 2026Q6_K11 GBfits41K~33~32~850
Gemma 4 31B Mar 2026IQ2_M13 GBfits, short context756~31~330
Gemma 4 26B-A4B Mar 2026IQ2_M10 GBfits92K~130~110~2,700
Gemma 3 27B Mar 2025IQ2_M11 GBfits14K~33~370
GLM
GLM-5.3 753B-A40B Aug 2026Q4_K_M462 GBdoes not fit
GLM-5.3 Flash 320B-A18B Aug 2026Q4_K_M197 GBdoes not fit
GLM-4.6 355B-A32B Sep 2025Q4_K_M221 GBdoes not fit
GLM-4.5 Air 106B-A12B Jul 2025Q4_K_M68 GBruns with 56 GB in system RAM~6.0~430
gpt-oss
gpt-oss-120b Aug 2025Q3_K_M59 GBruns with 48 GB in system RAM~17~15~1,000
gpt-oss-20b Aug 2025IQ2_M9 GBfits66K~120~82~2,800
Llama
Llama 4 Scout 109B-A17B Apr 2025Q3_K_M57 GBruns with 45 GB in system RAM~5.4~4.7~300
Llama 3.3 70B Dec 2024Q4_K_M47 GBruns with 36 GB in system RAM~1.1~72
Llama 3.1 8B Jul 2024Q8_011 GBfits15K~34~1,300
MiMo
MiMo V2.6 Pro 1.02T-A42B Sep 2026Q4_K_M624 GBdoes not fit
MiMo V2.6 Flash 309B-A15B Sep 2026Q4_K_M190 GBdoes not fit
Mistral
Mistral Medium 3.5 128B Mar 2026Q3_K_M67 GBruns with 56 GB in system RAM~0.7~40
Mistral Small 4 119B-A6.5B Mar 2026Q3_K_M65 GBruns with 53 GB in system RAM~11~780
Mistral Small 3.2 24B Jun 2025IQ2_M11 GBfits13K~34~430
Nemotron
Nemotron 3.5 Lightning 30B-A3B Aug 2026Q4_K_M21 GBruns with 9 GB in system RAM~32~31~1,500
Qwen
Qwen3.8 Flash-Next 125B-A6B Aug 2026Q3_K_M63 GBruns with 51 GB in system RAM~15~15~850
Qwen3.8 27B Aug 2026 (your pick)IQ2_M11 GBfits19K~34~380
Qwen3.6 27B Apr 2026IQ2_M11 GBfits19K~34~380
Qwen3.6 35B-A3B Apr 2026Q4_K_M23 GBruns with 12 GB in system RAM~33~31~1,700
Qwen3.5 9B Feb 2026Q8_011 GBfits28K~33~1,100
Qwen3 Coder 480B-A35B Jul 2025Q4_K_M296 GBdoes not fit
Qwen3 235B-A22B Jul 2025Q4_K_M146 GBdoes not fit
Qwen3 32B Apr 2025Q4_K_M24 GBruns with 12 GB in system RAM~3.1~2.1~160

Specifications

Memory
12 GB GDDR6X
Memory bandwidth
504 GB/s
Bus width
192-bit
Shader cores
5,888
Tensor cores
184
FP32
29.1 TFLOPS
FP16 (dense)
58 TFLOPS
FP8 (dense)
117 TFLOPS
Power
200 W
Form factor
Dual slot, full height
Interface
PCIe 4.0 x16
Architecture
Ada Lovelace
Launched
2023
MSRP
$599

Source: manufacturer specifications.

Embed the live price

A badge that always shows the current eBay low for the RTX 4070, delivered, and links back here. Free to use in a README, a forum signature or a blog post.

RTX 4070: live eBay low

HTML and Markdown snippets 
<a href="https://buyaigear.com/gpus/rtx-4070-12gb"><img src="https://buyaigear.com/badge/rtx-4070-12gb.svg" alt="RTX 4070: live eBay low" height="20"></a>
[![RTX 4070: live eBay low](https://buyaigear.com/badge/rtx-4070-12gb.svg)](https://buyaigear.com/gpus/rtx-4070-12gb)