buyaigear
Prices checked 7 min ago Sign in

NVIDIA · Ampere · 40 GB · Qwen3.8 27B · Q4_K_M at 8K context: ~59 tok/s

NVIDIA A100 40GB PCIe

The first PCIe A100: 40 GB of HBM2 at 1.5 TB/s with full Ampere tensor cores, in a card that drops into a workstation with airflow. A 32B model at Q8 fits; the 80 GB card costs about twice as much.

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

Lowest price right now

$5,599

eBay · free shipping · $139 per GB · observed 2026-11-14 18:37 UTC

Buy the cheapest on eBay, $5,599

free shipping · New · memorypartner_ltd 101,545 · 100.0%

See more on eBay

Or rent one from $0.40 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.

$5,598$5,599$5,599$5,600Nov 9Nov 14lowest $5,599

Where to buy

VendorPriceConditionObserved
eBay
cheapest active listing		 $5,599
free shipping		 new and used 2026-11-14 18:37 UTC Go to eBay

On eBay now

36 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 36 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 A100 40GB PCIe in the cloud right now, checked every hour.

CloudPer hourListed asAvailable
Vast.ai$0.40A100 PCIE3 verified offersRent on Vast.ai

At $0.40 an hour on Vast.ai, the $5,599 a used one costs on eBay buys about 14,000 hours of rental: 19 months around the clock, or 4.8 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 2026Q8_014 GBfits128K+~79~77~4,600
Gemma 4 31B Mar 2026Q8_035 GBfits71K~30~29~1,800
Gemma 4 26B-A4B Mar 2026Q8_028 GBfits128K+~130~120~14,000
Gemma 3 27B Mar 2025Q8_031 GBfits112K~34~32~2,000
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 2025IQ2_M39 GBfits12K~110~4,500
gpt-oss
gpt-oss-120b Aug 2025Q4_K_M73 GBruns with 34 GB in system RAM~24~22~5,400
gpt-oss-20b Aug 2025Q8_024 GBfits128K+~130~110~15,000
Llama
Llama 4 Scout 109B-A17B Apr 2025IQ2_M40 GBfits, short context7K~94~3,200
Llama 3.3 70B Dec 2024Q3_K_M39 GBfits11K~27~770
Llama 3.1 8B Jul 2024Q8_011 GBfits128K+~110~79~6,800
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 2026Q4_K_M83 GBruns with 43 GB in system RAM~0.9~0.8~210
Mistral Small 4 119B-A6.5B Mar 2026Q4_K_M79 GBruns with 39 GB in system RAM~16~10~4,200
Mistral Small 3.2 24B Jun 2025Q8_028 GBfits77K~38~33~2,300
Nemotron
Nemotron 3.5 Lightning 30B-A3B Aug 2026Q8_035 GBfits128K+~140~140~16,000
Qwen
Qwen3.8 Flash-Next 125B-A6B Aug 2026Q4_K_M78 GBruns with 38 GB in system RAM~20~19~4,500
Qwen3.8 27B Aug 2026 (your pick)Q8_031 GBfits128K+~35~33~2,000
Qwen3.6 27B Apr 2026Q8_031 GBfits128K+~35~33~2,000
Qwen3.6 35B-A3B Apr 2026Q8_039 GBfits66K~150~140~18,000
Qwen3.5 9B Feb 2026Q8_011 GBfits128K+~100~95~6,100
Qwen3 Coder 480B-A35B Jul 2025Q4_K_M296 GBdoes not fit
Qwen3 235B-A22B Jul 2025IQ2_M83 GBruns with 43 GB in system RAM~9.3~8.1~1,200
Qwen3 32B Apr 2025Q8_038 GBfits13K~27~1,700

Specifications

Memory
40 GB HBM2
Memory bandwidth
1,555 GB/s
Bus width
5120-bit
Shader cores
6,912
Tensor cores
432
FP32
19.5 TFLOPS
FP16 (dense)
312 TFLOPS
Power
250 W
Form factor
Dual slot, passive, needs chassis airflow
Interface
PCIe 4.0 x16
Architecture
Ampere
Launched
2020

Source: manufacturer specifications.

Embed the live price

A badge that always shows the current eBay low for the A100 40GB PCIe, delivered, and links back here. Free to use in a README, a forum signature or a blog post.

A100 40GB PCIe: live eBay low

HTML and Markdown snippets 
<a href="https://buyaigear.com/gpus/a100-40gb-pcie"><img src="https://buyaigear.com/badge/a100-40gb-pcie.svg" alt="A100 40GB PCIe: live eBay low" height="20"></a>
[![A100 40GB PCIe: live eBay low](https://buyaigear.com/badge/a100-40gb-pcie.svg)](https://buyaigear.com/gpus/a100-40gb-pcie)