How much VRAM does llama-3-firefunction-v2 need?

Q4_K_M is exactly 42,520,396,416 bytes (39.60 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is llama-3-firefunction-v2's KV cache?

10.00 GiB at 32K context with an f16 cache, computed per layer from the model's architecture. Quantizing the cache to q8_0 roughly halves it.

fireworks-ai · text

llama-3-firefunction-v2

fireworks-ai/llama-3-firefunction-v2

llama-3-firefunction-v2 at Q4_K_M is exactly 42,520,396,416 bytes (39.60 GiB / 42.52 GB) — an effective 4.821 bits per weight, not the nominal 4. Its KV cache at 32K is 10.00 GiB.

From the file· summed from 1 file(s)From the file· KV per layer

Parameters

70.6B

Architecture

llama

80 layers

Context

8,192

native (config.json)

License

llama3

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
IQ1_S	14.29 GiB	15,343,485,568	1.740	—	MaziyarPanahi
IQ1_M	15.60 GiB	16,751,198,848	1.899	—	MaziyarPanahi
IQ2_XS	19.69 GiB	21,142,110,848	2.397	—	MaziyarPanahi
Q2_K	24.56 GiB	26,375,111,296	2.991	—	MaziyarPanahi
IQ3_XS	27.29 GiB	29,307,732,608	3.323	—	MaziyarPanahi
Q3_K_S	28.79 GiB	30,912,053,888	3.505	—	MaziyarPanahi
Q3_K_M	31.91 GiB	34,267,497,088	3.886	723	MaziyarPanahi
Q3_K_L	34.59 GiB	37,140,595,328	4.211	—	MaziyarPanahi
IQ4_XS	35.30 GiB	37,902,664,320	4.298	723	MaziyarPanahi
Q4_K_S	37.58 GiB	40,347,222,656	4.575	—	MaziyarPanahi
Q4_K_M	39.60 GiB	42,520,396,416	4.821	723	MaziyarPanahi
Q5_K_S	45.32 GiB	48,657,449,600	5.517	—	MaziyarPanahi
Q5_K_M	46.52 GiB	49,949,819,520	5.664	723	MaziyarPanahi
Q6_K6 shards	53.91 GiB	57,888,146,752	6.564	—	MaziyarPanahi
Q8_06 shards	69.83 GiB	74,975,053,120	8.501	—	MaziyarPanahi

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	1.25 GiB	1.25 GiB	—	80 / 0 / 0
8,192	2.50 GiB	2.50 GiB	—	80 / 0 / 0
16,384	5.00 GiB	5.00 GiB	—	80 / 0 / 0
32,768	10.00 GiB	10.00 GiB	—	80 / 0 / 0
65,536	20.00 GiB	20.00 GiB	—	80 / 0 / 0
131,072	40.00 GiB	40.00 GiB	—	80 / 0 / 0

Compare with

same modality, comparable size

vs Qwen3-Coder-30B-A3B-Instruct vs Qwen3.6-27B vs Qwen3.8-27B vs gemma-4-12B-it-qat-q4_0-unquantized vs gemma-4-E4B-it vs Qwen3-30B-A3B-Thinking-2507 All alternatives →Will it fit on my card? →

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 36.96 GiB. The real file is 39.60 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

128

Hidden size

8192

Vocab

128,256

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does llama-3-firefunction-v2 need?: Q4_K_M is exactly 42,520,396,416 bytes (39.60 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is llama-3-firefunction-v2's KV cache?: 10.00 GiB at 32K context with an f16 cache, computed per layer. Quantizing the cache to q8_0 roughly halves it, which is often the difference between a context length fitting and not.
Which quantization of llama-3-firefunction-v2 should I use?: Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.