what/llm/can/i/run
gpus/nvidia/a100-40gb

A100 40GB

Released 2020-05-14 · Street $8,000
vram
40 GB
discrete
bandwidth
1,555
GB/s
models fit
16
top · Q8_0
power
400 W
high

Spec sheet · top pick

Hardware spec
A100 40GB
NVIDIA · Ampere · sm_80
vram
40 GB
bandwidth
1,555 GB/s
tdp
400 W
released
2020-05-14
msrp
$11,000
street
~$8,000
★ Sweet spot for this card
gemma 3 27b
Q8_0 · GGUF · 28.9 GB
bench50.6 · conf A
vram28.9 / 40 GB
speed54 tok/s
Best benchmark per GB of VRAM in your tier. Q8_0wastes <3% quality vs FP16, leaves headroom for 32k context.
Full model inspect →

Top 10 models on this GPU

#modelquantvramtok/sbench ↓fit
01gemma 3 27bQ8_028.9 GB5450.6Great
02Qwen2.5 Coder 32BQ8_034.1 GB4648.8Great
03Gemma 2 27BQ8_028.9 GB5446.2Great
04Qwen3 32BQ8_034.1 GB4640.0Great
05deepseek r1 distill llama 70bQ4_K_M37.3 GB4251.7Ok
06Gemma 2 9BFP1619.4 GB8035.0Great
07Gemma 3 12BFP1625.7 GB6131.3Great
08deepseek r1 distill qwen 32bQ8_034.1 GB4633.3Great
09Gemma 3 4BFP168.9 GB17515.1Great
10Mixtral 8x7B v0.1Q6_K37.3 GB4213.3Ok
Showing 10 of 16 models that fit · see full ranked list →

Upgrade options

★ Best value
+A6000 Pro
48 GB · ~$4,200 · +$-3,800 over A100 40GB
Unlocks +1 models not in current rank
See A6000 Pro
Bigger tier
+RTX 6000 Ada
48 GB · ~$7,000 · +$-1,000 over A100 40GB
Unlocks +1 models not in current rank
See RTX 6000 Ada
Premium tier
+H100 PCIe
80 GB · ~$28,000 · +$20,000 over A100 40GB
Unlocks +2 models not in current rank
See H100 PCIe

Also compare

RTX 5090
32 GB · NVIDIA · 70% cheaper
RTX 4090
24 GB · NVIDIA · 76% cheaper
RTX 3090 Ti
24 GB · NVIDIA · 88% cheaper
RTX 3090
24 GB · NVIDIA · 89% cheaper
Not running a A100 40GB?
Tell us what you actually have — we'll re-rank everything for your machine.
Change my machine