How much VRAM does Llama-4-Scout-17B-16E-Instruct-4bit need?

Its published quantizations range from 20.95 GiB to 37.96 GiB in weights alone, before KV cache and runtime overhead.

Llama-4-Scout-17B-16E-Instruct-4bit on Apple M3 Pro

Can I run Llama-4-Scout-17B-16E-Instruct-4bit on a Apple M3 Pro?

Q: Can I run Llama-4-Scout-17B-16E-Instruct-4bit on a Apple M3 Pro?

No. No published quantization of Llama-4-Scout-17B-16E-Instruct-4bit fits a Apple M3 Pro at any context with q4_0 KV cache.

Not at these settings. No indexed quantization of Llama-4-Scout-17B-16E-Instruct-4bit fits Apple M3 Pro at any context we compute, with q4_0 KV. The smallest shipped quantization is 20.95 GiB in weights alone, against 12.56 GiB usable. CPU offload can still run it, slowly.

From the file· weights summed from filesFrom the file· KV computed per layerPredicted· speed and compute buffer

KV cache dtypef16 q8_0 q4_0— 18 GB card, 12.56 GiB usable after overhead

Every quantization at every context

total memory required; green fits, red does not

Quant	Weights●	4K◐	8K◐	16K◐	32K◐	64K◐	128K◐
F16	37.96 GiB	38.8	39.0	39.4	40.2	41.9	45.3
Q8_0	26.16 GiB	27.0	27.2	27.6	28.4	30.1	33.5
Q4_K_M	20.95 GiB	21.7	21.9	22.4	23.2	24.9	28.3

●From the file◐Predictedwhat these mean

Figures are GiB of total memory: weights plus KV cache plus compute buffer and backend overhead. Weights and KV are near-exact; the overhead term is modeled. Hover any cell for the breakdown.

Why other calculators disagree

A parameters × bits ÷ 8 estimate ignores two things that dominate at long context. First, the weights themselves are not the nominal rate — quantizations are mixtures, so the real file is consistently larger than the label implies. Second, most of this model's layers cache only a 8,192-token window rather than the full context.