aliteq.

A100 PCIe: 80 GB, what it runs and what it costs

Specs checked · Rental prices from 49m ago · Aliteq Team · How we compute this

The short answer

The NVIDIA A100 80GB PCIe has 80 GB of VRAM and 1935 GB/s of memory bandwidth. At Q4_K_M quality or better with an 8K context it runs 40 of the open models we track; the largest that fits with room to spare is Qwen3.5 122B-A10B (MoE). The cheapest rental right now is $1.790/hr (Runpod, on demand).

VRAM
80 GB
Bandwidth
1935 GB/s
Board power
300 W
Launched
—

Source: NVIDIA specification · bandwidth not yet confirmed by a second source, so we don't estimate speed from it

Buy it or rent it

ProviderTypeFrom, per hourRent
RunpodOn demand$1.790/hrRunpod

Lowest price per provider in our hourly snapshot (49m ago). Rent buttons are referral links. Every card side by side: cloud GPU prices.

Which AI models run on the A100 PCIe

Computed by our VRAM engine for an 8K context: the best quality each model fits at in 80 GB. Speeds are upper-bound estimates from memory bandwidth, not benchmarks.

ModelBest fitVRAM used~tok/s
Qwen3.5 122B-A10B (MoE)Q4_K_M72 GB89%—
NVIDIA Nemotron 3 Super 120B-A12B (MoE)Q4_K_M71 GB88%—
gpt-oss 120B (MoE)Q5_K_M79 GB98% · tight—
Llama 4 Scout 17B-16E (MoE)Q5_K_M73 GB91% · tight—
GLM-4.5-Air (MoE)Q5_K_M72 GB90% · tight—
Qwen3-Next 80B-A3B (MoE)Q6_K62 GB78%—
Qwen3-Coder-Next 80B-A3B (MoE)Q8_080 GB100% · tight—
DeepSeek R1 Distill 70BQ8_073 GB91% · tight—
Mixtral 8x7B (MoE)Q8_048 GB60%—
Qwen3.6 35B-A3B (MoE)F1668 GB85%—
Ornith-1.5 35B-A3B (MoE)F1668 GB85%—
Qwen3.5 35B-A3B (MoE)F1668 GB85%—

Fits only below Q4 quality: Qwen3.8-Flash-Next (MoE) (Q2_K).

Too big for 80 GB (rent a bigger card instead): what it costs to run MiniMax-M2.5 (MoE), what it costs to run Qwen3 235B-A22B (MoE), what it costs to run MiniMax-M3 (MoE), what it costs to run GLM-5.3 (MoE).

Check a model and context of your own in the VRAM calculator.

AI video on the A100 PCIe

Alibaba's Wan 2.2 is an open video model you can run yourself in ComfyUI. With 80 GB the A100 PCIe fits:

  • Wan 2.2 14B at 720p without offloading (80 GB and up, per Runpod's Wan 2.2 guide)
  • Wan 2.2 14B in fp16, or in fp8 at 720p (48 GB and up, per Runpod's Wan 2.2 guide)
  • Wan 2.2 14B at 640x640 with fp8 weights (24 GB and up, per Runpod's Wan 2.2 guide)
  • Wan 2.2 5B, with ComfyUI's native offloading (8 GB and up, per ComfyUI's Wan 2.2 docs)

What each setup needs, the cheapest card to rent and what one clip costs: cloud GPUs for ComfyUI with Flux and Wan 2.2. Against an app or an API: AI video cost per clip, app vs API vs rented GPU. Rather not run a model yourself? Hosted video models such as Seedance, Veo and Hailuo are compared in Kling AI alternatives, priced per clip.

Our guides on the A100 PCIe

Cards to compare it with

Every card by VRAM tier: all GPUs for gaming and local AI.

A100 PCIe: common questions

How much VRAM does the A100 PCIe have?
80 GB, with 1935 GB/s of memory bandwidth, per the maker's specification.
What is the biggest AI model the A100 PCIe can run?
Of the open models we track, Qwen3.5 122B-A10B (MoE) is the largest that fits at Q4_K_M or better with an 8K context and room to spare (about 72 GB of 80 GB). Bigger models only fit at lower-quality quants or need more memory.
How much does it cost to rent an A100 PCIe?
From $1.790/hr on Runpod (on demand), the lowest price in our hourly snapshot of Nov 7, 2026. Prices move every hour.
Is 80 GB enough?
Datacenter GPUs and unified-memory machines; mostly rented by the hour.