dexiio

What hardware runs Llama 3.3 70B?

Llama 3.3 70B is a 70.55B-parameter dense model (llama3.3 license). Still the dense 70B reference. Prose depth the MoEs haven't matched; the hardware bill is the price of admission.

All figures assume an f16 KV cache, a 0.6 GB display reserve on the GPU, and 64 GB of DDR5 system RAM for the offload tiers. Tune these in the calculator.

Minimum VRAM by quant and context

QuantFile size8K16K32K64K128K
Q2_K *24.6 GB28.9 GB31.5 GB36.6 GB46.9 GB67.4 GB
Q3_K_M31.9 GB36.3 GB38.8 GB44 GB54.2 GB74.7 GB
Q4_K_M39.6 GB44 GB46.5 GB51.7 GB61.9 GB82.4 GB
Q5_K_M46.5 GB50.9 GB53.4 GB58.6 GB68.8 GB89.3 GB
Q6_K53.9 GB58.3 GB60.8 GB66 GB76.2 GB96.7 GB
Q8_069.8 GB74.2 GB76.8 GB81.9 GB92.1 GB112.6 GB
IQ4_XS35.3 GB39.7 GB42.3 GB47.4 GB57.6 GB78.1 GB

Full-GPU figures: weights + f16 KV cache + overhead. * below our recommended floor of Q3_K_M.

GPU compatibility

GPU8K16K32K64K128K
NVIDIA GeForce RTX 5090IQ4_XSIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 5080IQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 5070 TiIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 5070IQ4_XSIQ4_XSQ4_K_M
NVIDIA GeForce RTX 5060 Ti 16GBIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 5060IQ4_XSQ4_K_MQ4_K_M
NVIDIA GeForce RTX 4090IQ4_XSIQ4_XSIQ4_XSQ3_K_M
NVIDIA GeForce RTX 4080 SUPERIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 4070 Ti SUPERIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 4070IQ4_XSIQ4_XSQ4_K_M
NVIDIA GeForce RTX 4060 Ti 16GBIQ4_XSIQ4_XSIQ4_XS
NVIDIA GeForce RTX 4060IQ4_XSQ4_K_MQ4_K_M
NVIDIA GeForce RTX 3090IQ4_XSIQ4_XSIQ4_XSQ3_K_M
NVIDIA GeForce RTX 3080 10GBIQ4_XSIQ4_XSQ4_K_M
NVIDIA GeForce RTX 3070IQ4_XSQ4_K_MQ4_K_M
NVIDIA GeForce RTX 3060 TiIQ4_XSQ4_K_MQ4_K_M
NVIDIA GeForce RTX 3060 12GBIQ4_XSIQ4_XSQ4_K_M
NVIDIA GeForce RTX 2080 TiIQ4_XSIQ4_XSQ4_K_M
NVIDIA GeForce GTX 1080 TiIQ4_XSIQ4_XSQ4_K_M
AMD Radeon RX 9070 XTIQ4_XSIQ4_XSIQ4_XS
AMD Radeon RX 7900 XTXIQ4_XSIQ4_XSIQ4_XSQ3_K_M
AMD Radeon RX 7900 XTIQ4_XSIQ4_XSIQ4_XS
AMD Radeon RX 7800 XTIQ4_XSIQ4_XSIQ4_XS
AMD Radeon RX 7600 XTIQ4_XSIQ4_XSIQ4_XS
AMD Radeon RX 6800 XTIQ4_XSIQ4_XSIQ4_XS
AMD Radeon RX 6700 XTIQ4_XSIQ4_XSQ4_K_M
Intel Arc B580IQ4_XSIQ4_XSQ4_K_M
Intel Arc A770 16GBIQ4_XSIQ4_XSIQ4_XS
Apple M2 (16GB)IQ4_XSIQ4_XSQ4_K_M
Apple M2 Pro (32GB)IQ4_XSIQ4_XSIQ4_XSQ3_K_M
Apple M2 Max (64GB)Q4_K_MQ4_K_MIQ4_XSIQ4_XSIQ4_XS
Apple M2 Ultra (128GB)Q8_0Q8_0Q8_0Q8_0Q5_K_M
Apple M3 (16GB)IQ4_XSIQ4_XSQ4_K_M
Apple M3 Pro (36GB)IQ4_XSIQ4_XSIQ4_XSIQ4_XS
Apple M3 Max (64GB)Q4_K_MQ4_K_MIQ4_XSIQ4_XSIQ4_XS
Apple M3 Ultra (96GB)Q6_KQ6_KQ6_KQ5_K_MIQ4_XS
Apple M4 (16GB)IQ4_XSIQ4_XSQ4_K_M
Apple M4 Pro (48GB)IQ4_XSIQ4_XSIQ4_XSIQ4_XS
Apple M4 Max (64GB)Q4_K_MQ4_K_MIQ4_XSIQ4_XSIQ4_XS
CPU only / integrated graphicsQ5_K_MQ4_K_MQ4_K_M

Fits on GPUExpert offloadPartial offloadCPU only

Quant guidance

Our floor for Llama 3.3 70B is Q3_K_M — below that, quality degrades faster than the VRAM savings are worth (large models tolerate low bpw better than small ones, which is why the floor is lower here). Prefer the highest quant that still lands "Fits on GPU" at your context length in the table above.

Recommended run command

Q3_K_M at 32K on a Apple M2 Max (64GB)-class GPU (Fits on GPU):

llama-server -m Llama-3.3-70B-Instruct-Q3_K_M.gguf -c 32768 --flash-attn -ngl 99

Frequently asked questions

How much VRAM does Llama 3.3 70B need?

At Q3_K_M and 32K context, Llama 3.3 70B needs about 44 GB of VRAM to run fully on GPU (weights + KV cache + overhead).

What is the smallest GPU that can run Llama 3.3 70B?

The Apple M2 Max (64GB) (48 GB) is the smallest GPU in our set that runs Llama 3.3 70B well at 32K context.

What quantization should I use for Llama 3.3 70B?

We recommend Q3_K_M or higher. Q4_K_M weighs 39.6 GB (4.82 bits/weight); going below Q3_K_M costs noticeable quality on a model this size.

How long a context can Llama 3.3 70B handle?

Llama 3.3 70B supports up to 128K tokens. KV cache grows linearly with context.

Can I run Llama 3.3 70B without a GPU?

Yes, at reduced speed: on CPU with 64 GB of DDR5 it manages roughly 1 tokens/sec at 8K context (Q5_K_M).

Related guides