Llama 3.2 90B Vision Instruct — VRAM Requirements
Written by Jakub Rusinowski · Last updated August 15, 2026
How much GPU VRAM you need to run Llama 3.2 Family Llama 3.2 90B Vision Instruct by Meta locally, a 88.8B-parameter model. Figures are quantized weights + KV cache + framework overhead, computed from the model's parameter count and published architecture — not a throughput model. See /en/methodology.
Llama 3.2 90B Vision Instruct needs about 58 GB VRAM at Q4_K_M.
VRAM by Quantization
| Quant | Bits/weight | Weights | Total VRAM |
|---|---|---|---|
| Q2_K | 2.63 | 29.2 GB | 33.3 GB |
| Q3_K_M | 3.41 | 37.9 GB | 42.0 GB |
| Q4_K_M | 4.83 | 53.6 GB | 57.8 GB |
| Q5_K_M | 5.67 | 62.9 GB | 67.1 GB |
| Q6_K | 6.56 | 72.8 GB | 77.0 GB |
| Q8_0 | 8.50 | 94.3 GB | 98.5 GB |
| F16 | 16.00 | 177.6 GB | 181.8 GB |
Switch quantization in the interactive calculator, or see the full Llama 3.2 Family model page.
or compare on Vast.ai from $0.77/hr (typical low · varies)
As an Amazon Associate we earn from qualifying purchases. Cloud GPU links are referral links — we may earn a commission at no extra cost to you.
Add this badge to your model card
Model creators: paste this into your Hugging Face model card README to link readers straight to this VRAM breakdown.
[](https://llmconfigurator.com/en/vram-calculator/llama-3.2-90b-vision?utm_source=badge&utm_medium=referral&utm_campaign=readme_badge&utm_content=llama-3.2-90b-vision)Estimates only — actual VRAM varies with context length, batch size, runtime and KV-cache settings.