Llama 3.2 Vision (11B) has 11 billion parameters. At standard 4-bit quantization with 8K context, it needs roughly 9.4 GB of VRAM — weights plus cache and runtime overhead.

VRAM by quantization

PrecisionWeightsCache/BufferTotal VRAM
2-bit (IQ2_XXS)3.5 GB1.3 GB6.8 GB
4-bit (Q4_K_M)6.1 GB1.3 GB9.4 GB
8-bit (Q8_0)11.6 GB1.3 GB14.9 GB
16-bit (FP16)22.0 GB1.3 GB25.3 GB

Which GPU can run Llama 3.2 Vision (11B) (at 4-bit)?

GPU classVRAMLlama 3.2 Vision (11B) (9.4 GB)
8 GB · RTX 5060 / 40608 GBTight
12 GB · RTX 5070 / 306012 GBFits
16 GB · RTX 5070 Ti / 408016 GBFits
24 GB · RTX 4090 / 309024 GBFits
32 GB · RTX 509032 GBFits
48 GB · 2×24 / RTX 6000 Ada48 GBFits
128 GB · M-series / RTX Spark128 GBFits

Meta's multi-modal conversational model.

Get the exact number for your setup
Pick your model, quantization, and context length — the calculator shows the full VRAM math and tells you precisely which hardware fits.
Open the Local AI Calculator
Related guides
Best GPU for Llama 3 70B How Much VRAM for DeepSeek-R1 Q4 vs Q8 Quantization Explained Apple Silicon for Local AI RTX Spark: 128GB Unified Memory

VRAM figures are reproducible estimates (weights + KV cache + overhead) and vary by runtime and quant format. Data current as of 2026-06-15.