RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run Llama 3.2 1B Instruct locally?

Llama 3.2 1B Instruct by meta-llama has 1.2B parameters. At 4-bit with an 8K context it needs about 1.6 GB of GPU memory. It fits on 65 of 65 GPUs and Macs we track. See all models that fit in 8 GB of VRAM.

1.6 GBQ4_K_M · 8K context
RTX 4060smallest consumer GPU that fits
M1 16GBsmallest Mac that fits
128Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Llama-3.2-1B-Instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough1.3 GB1.4 GB2.2 GB5.2 GB
3-bit (Q3_K_M) •usable1.4 GB1.5 GB2.3 GB5.3 GB
4-bit (Q4_K_M) •good, most popular1.5 GB1.6 GB2.4 GB5.4 GB
5-bit (Q5_K_M) •very good1.6 GB1.7 GB2.5 GB5.5 GB
6-bit (Q6_K) •near lossless1.7 GB1.8 GB2.6 GB5.6 GB
8-bit (Q8_0) •effectively lossless2.0 GB2.1 GB2.9 GB5.9 GB
16-bit (F16/BF16) •original weights5.4 GB5.5 GB6.3 GB9.3 GB

Which GPUs and Macs run Llama 3.2 1B Instruct?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBFitsF165.5 GB73details
RTX 4060 Ti 16GB16 GBFitsF165.5 GB58details
RTX 309024 GBFitsF165.5 GB189details
RTX 409024 GBFitsF165.5 GB203details
RTX 509032 GBFitsF165.5 GB361details
RTX 5070 Ti16 GBFitsF165.5 GB181details
RX 7900 XTX24 GBFitsF165.5 GB194details
Mac M4 Pro 48GB48 GBFitsF165.5 GB55details
Mac M4 Max 128GB128 GBFitsF165.5 GB110details
Mac M3 Ultra 512GB512 GBFitsF165.5 GB165details
DGX Spark128 GBFitsF165.5 GB55details
RTX 407012 GBFitsF165.5 GB102details
RTX 5060 Ti 16GB16 GBFitsF165.5 GB90details
RTX 40608 GBFitsF165.5 GB55details
Arc B58012 GBFitsF165.5 GB92details
RTX 3060 Ti8 GBFitsF165.5 GB90details
RTX 30708 GBFitsF165.5 GB90details
RTX 3080 10GB10 GBFitsF165.5 GB153details
RTX 3080 12GB12 GBFitsF165.5 GB184details
RTX 3090 Ti24 GBFitsF165.5 GB203details
RTX 4060 Ti 8GB8 GBFitsF165.5 GB58details
RTX 4070 Super12 GBFitsF165.5 GB102details
RTX 4070 Ti12 GBFitsF165.5 GB102details
RTX 4070 Ti Super16 GBFitsF165.5 GB136details
RTX 408016 GBFitsF165.5 GB145details
RTX 4080 Super16 GBFitsF165.5 GB148details
RTX 50608 GBFitsF165.5 GB90details
RTX 507012 GBFitsF165.5 GB136details
RTX 508016 GBFitsF165.5 GB194details
2x RTX 30902×24 GBFitsF165.5 GB189details
2x RTX 40902×24 GBFitsF165.5 GB203details
2x RTX 50902×32 GBFitsF165.5 GB361details
RTX A600048 GBFitsF165.5 GB155details
RTX 6000 Ada48 GBFitsF165.5 GB194details
RTX PRO 6000 Blackwell96 GBFitsF165.5 GB361details
L40S48 GBFitsF165.5 GB174details
A100 40GB40 GBFitsF165.5 GB314details
A100 80GB80 GBFitsF165.5 GB411details
H100 PCIe80 GBFitsF165.5 GB403details
H100 SXM80 GBFitsF165.5 GB676details
H200141 GBFitsF165.5 GB968details
RX 76008 GBFitsF165.5 GB58details
RX 7700 XT12 GBFitsF165.5 GB87details
RX 7800 XT16 GBFitsF165.5 GB126details
RX 7900 GRE16 GBFitsF165.5 GB116details
RX 7900 XT20 GBFitsF165.5 GB161details
RX 907016 GBFitsF165.5 GB129details
RX 9070 XT16 GBFitsF165.5 GB129details
Ryzen AI Max+ 395 128GB128 GBFitsF165.5 GB52details
Arc A770 16GB16 GBFitsF165.5 GB113details
Mac M1 16GB16 GBFitsF165.5 GB14details
Mac M2 24GB24 GBFitsF165.5 GB20details
Mac M3 24GB24 GBFitsF165.5 GB20details
Mac M4 16GB16 GBFitsF165.5 GB24details
Mac M4 32GB32 GBFitsF165.5 GB24details
Mac M1 Pro 32GB32 GBFitsF165.5 GB40details
Mac M3 Pro 36GB36 GBFitsF165.5 GB30details
Mac M4 Pro 24GB24 GBFitsF165.5 GB55details
Mac M4 Pro 64GB64 GBFitsF165.5 GB55details
Mac M1 Max 64GB64 GBFitsF165.5 GB81details
Mac M2 Max 96GB96 GBFitsF165.5 GB81details
Mac M3 Max 128GB128 GBFitsF165.5 GB81details
Mac M4 Max 64GB64 GBFitsF165.5 GB110details
Mac M2 Ultra 192GB192 GBFitsF165.5 GB161details
Mac M3 Ultra 256GB256 GBFitsF165.5 GB165details

Prefer an API?

Meta: Llama 3.2 1B Instruct is on OpenRouter at $0.027 per 1M input tokens and $0.201 per 1M output tokens. Price details and history →

Questions

How much VRAM does Llama 3.2 1B Instruct need?

About 1.6 GB at 4-bit (Q4_K_M) with an 8K context: 0.8 GB of weights (real GGUF file size), 0.3 GB of KV cache and 0.6 GB of runtime overhead. At 8-bit it needs about 2.1 GB.

What is the smallest GPU that runs Llama 3.2 1B Instruct?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 4060 (8 GB).

Can I run Llama 3.2 1B Instruct on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M1 16GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

Llama 3.2 1B

1.2B · 1.8 GB at 4-bit

Llama 3.2 3B Instruct

3.2B · 3.4 GB at 4-bit

Llama 3.2 3B

3.2B · 3.6 GB at 4-bit

LFM2.5 1.2B Instruct

1.2B · 1.3 GB at 4-bit

deepseek coder 1.3b instruct

1.3B · 2.8 GB at 4-bit

HunyuanOCR

1.1B · 2.0 GB at 4-bit

Model facts

Hugging Face repometa-llama/Llama-3.2-1B-Instruct
Parameters1,235,814,400 (1.2B)
Layers / hidden size16 / 2048
AttentionGQA; 32 heads, 8 KV heads, head dim 64
KV cache per token (FP16)32.0 KiB
Max context131,072 tokens
Licensellama3.2 (gated: accept terms on Hugging Face)
Released on HFSep 18, 2024
Downloads (30d) / likes1,313,473 / 1,989

Sources: Hugging Face model API · config.json (unsloth/Llama-3.2-1B-Instruct-GGUF) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.