what/llm/can/i/run
gpus/nvidia/rtx-4090

RTX 4090

Released 2022-10-12 · Street $1,899
vram
24 GB
discrete
bandwidth
1,008
GB/s
models fit
13
top · Q6_K
power
450 W
high

Spec sheet · top pick

Hardware spec
RTX 4090
NVIDIA · Ada Lovelace · sm_89
vram
24 GB
bandwidth
1,008 GB/s
tdp
450 W
released
2022-10-12
msrp
$1,599
street
~$1,899
★ Sweet spot for this card
gemma 3 27b
Q6_K · GGUF · 21.8 GB
bench50.6 · conf A
vram21.8 / 24 GB
speed46 tok/s
Best benchmark per GB of VRAM in your tier. Q6_Kwastes <3% quality vs FP16, leaves headroom for 32k context.
Full model inspect →

Top 10 models on this GPU

#modelquantvramtok/sbench ↓fit
01gemma 3 27bQ6_K21.8 GB4650.6Ok
02Qwen2.5 Coder 32BQ5_K_M21.5 GB4748.8Ok
03Gemma 2 27BQ6_K21.8 GB4646.2Ok
04Qwen3 32BQ5_K_M21.5 GB4740.0Ok
05Gemma 2 9BFP1619.4 GB5235.0Great
06Gemma 3 12BQ8_013.1 GB7731.3Great
07deepseek r1 distill qwen 32bQ5_K_M21.5 GB4733.3Ok
08Gemma 3 4BFP168.9 GB11315.1Great
09Mixtral 8x7B v0.1Q2_K18.9 GB5313.3Great
10Gemma 2 2bFP164.7 GB2140.0Great
Showing 10 of 13 models that fit · see full ranked list →

Upgrade options

★ Best value
+RTX 5090
32 GB · ~$2,399 · +$500 over RTX 4090
Unlocks +3 models not in current rank
See RTX 5090
Bigger tier
+A6000 Pro
48 GB · ~$4,200 · +$2,301 over RTX 4090
Unlocks +4 models not in current rank
See A6000 Pro
Premium tier
+RTX 6000 Ada
48 GB · ~$7,000 · +$5,101 over RTX 4090
Unlocks +4 models not in current rank
See RTX 6000 Ada

Also compare

RTX 5090
32 GB · NVIDIA · 26% more
RTX 5080
16 GB · NVIDIA · 42% cheaper
RTX 5070 Ti
16 GB · NVIDIA · 55% cheaper
RTX 5070
12 GB · NVIDIA · 68% cheaper
Not running a RTX 4090?
Tell us what you actually have — we'll re-rank everything for your machine.
Change my machine