How much VRAM does gemma-2-9b-it-bnb-4bit need?

Q8_0 is exactly 9,827,148,704 bytes (9.15 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is gemma-2-9b-it-bnb-4bit's KV cache?

5.99 GiB at 32K context with an f16 cache, computed per layer from the model's architecture — most of its layers use sliding-window attention, so a flat formula overstates this substantially. Quantizing the cache to q8_0 roughly halves it.

unsloth · text

gemma-2-9b-it-bnb-4bit

unsloth/gemma-2-9b-it-bnb-4bit

gemma-2-9b-it-bnb-4bit at Q8_0 is exactly 9,827,148,704 bytes (9.15 GiB / 9.83 GB) — an effective 8.273 bits per weight, not the nominal 8. Its KV cache at 32K is 5.99 GiB, not the 10.50 GiB a flat formula predicts.

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

Parameters

9.5B

Architecture

gemma2

42 layers

Context

8,192

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q8_0	9.15 GiB	9,827,148,704	8.273	—	twright8

KV cache by context

computed per layer — this model uses sliding-window attention

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	1.31 GiB	1.31 GiB	—	21 / 21 / 0
8,192	2.05 GiB	2.63 GiB	1.28×	21 / 21 / 0
16,384	3.36 GiB	5.25 GiB	1.56×	21 / 21 / 0
32,768	5.99 GiB	10.50 GiB	1.75×	21 / 21 / 0
65,536	11.24 GiB	21.00 GiB	1.87×	21 / 21 / 0
131,072	21.74 GiB	42.00 GiB	1.93×	21 / 21 / 0

21 of 42 layers cache only a 4,096-token window rather than the full context, on a period of 2. Figures assume the default configuration; --swa-full disables the saving entirely.

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 Q8_0 at roughly 4.98 GiB. The real file is 9.15 GiB, because a quantization is a mixture and some tensors are always kept at higher precision. The larger discrepancy is the cache: a flat formula gives 10.50 GiB at 32K context where the real figure is 5.99 GiB, because most of this model's layers cache a fixed window rather than the whole context.

Architecture

from config.json

Layers

Attention heads

KV heads

Head dim

256

Hidden size

3584

Vocab

256,000

Sliding window

4096

SWA period

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does gemma-2-9b-it-bnb-4bit need?: Q8_0 is exactly 9,827,148,704 bytes (9.15 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is gemma-2-9b-it-bnb-4bit's KV cache?: 5.99 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 gemma-2-9b-it-bnb-4bit 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.