What GPU do I need to run meta-llama/Llama-2-13b-chat-hf?
13.0B parameters, published in F16. View on Hugging FaceGated
Set up Llama-2-13b-chat-hfLlama-2-13b-chat-hf is published by meta-llama on Hugging Face, with 116,347 downloads and 1,124 likes to date. It's a LlamaForCausalLM model built for text-generation, published natively in F16, and gated: you'll need to accept the model's terms on Hugging Face before downloading weights.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision, plus a fixed overhead for KV-cache, activations, and fragmentation. Full formula and assumptions: methodology.
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
Cheapest way to run Llama-2-13b-chat-hf at its published (F16) precision: 1× V100 on simplepod, at $0.170/hr per GPU ($0.170/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Llama-2-13b-chat-hf: common questions
Does Llama-2-13b-chat-hf fit on a 32 GB GPU?
Yes. At FP16 it needs 29.1 GB of VRAM, so a 32 GB card holds it with 2.9 GB to spare. A 24 GB card is not enough for it at FP16.
Do I need approval to download Llama-2-13b-chat-hf?
Yes. meta-llama gates this repository on Hugging Face, so you have to accept its terms with a signed-in account before the weights will download. It does not change the 29.1 GB the model needs once you have them.
What is the least VRAM Llama-2-13b-chat-hf can run in?
7.3 GB, at INT4 (quantized), which fits an 8 GB card, against 29.1 GB at FP16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.
Does quantizing Llama-2-13b-chat-hf lower the GPU bill?
Yes. At FP16 the cheapest live fit is one V100 on simplepod at $0.170/hr. At INT4 (quantized) it drops to one RTX 3060 on simplepod at $0.080/hr, provided a quantized checkpoint exists for it.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More meta-llama models
- Llama-3.2-1B-Instruct (1.2B, BF16)
- Llama-3.1-8B-Instruct (8.0B, BF16)
- Meta-Llama-3-8B-Instruct (8.0B, BF16)
- Llama-3.2-3B-Instruct (3.2B, BF16)
- Llama-3.2-1B (1.2B, BF16)
- Llama-2-7b-hf (6.7B, F16)
