How much VRAM does Qwen3-0.6B-Base need?

Q4_K_M is exactly 396,476,256 bytes (0.37 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is Qwen3-0.6B-Base's KV cache?

3.50 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.

Qwen · text

Qwen3-0.6B-Base

Qwen/Qwen3-0.6B-Base

Qwen3-0.6B-Base at Q4_K_M is exactly 396,476,256 bytes (0.37 GiB / 0.40 GB) — an effective 5.321 bits per weight, not the nominal 4. Its KV cache at 32K is 3.50 GiB.

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

Parameters

596M

Architecture

qwen3

28 layers

Context

32,768

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
TQ1_0	0.29 GiB	314,252,544	4.218	—	Majipa
TQ2_0	0.31 GiB	334,896,384	4.495	—	Majipa
Q2_K	0.32 GiB	347,288,832	4.661	—	Majipa
Q3_K_S	0.36 GiB	389,927,168	5.234	—	Majipa
IQ3_S	0.36 GiB	389,927,168	5.234	—	Majipa
Q4_K_M	0.37 GiB	396,476,256	5.321	311	iyanello
IQ3_M	0.38 GiB	402,878,720	5.407	—	Majipa
Q3_K_M	0.39 GiB	413,978,880	5.556	311	Majipa
Q3_K_L	0.41 GiB	435,343,616	5.843	—	Majipa
IQ4_XS	0.42 GiB	451,931,392	6.066	311	Majipa
Q4_0	0.44 GiB	469,081,344	6.296	311	Majipa
IQ4_NL	0.44 GiB	470,260,992	6.312	—	Majipa
Q4_K_S	0.44 GiB	470,785,280	6.319	—	Majipa
Q4_K_M	0.45 GiB	484,220,160	6.499	311	Majipa
Q4_1	0.47 GiB	506,330,368	6.796	—	Majipa
Q5_K_S	0.51 GiB	543,579,392	7.296	—	Majipa
Q5_0	0.51 GiB	543,579,392	7.296	—	Majipa
Q5_K_M	0.51 GiB	551,378,176	7.400	311	Majipa
Q5_1	0.54 GiB	580,828,416	7.796	—	Majipa
Q6_K	0.58 GiB	622,733,568	8.358	311	Majipa
Q8_0	0.60 GiB	639,150,592	8.579	310	Qwen
Q8_0	0.60 GiB	639,153,312	8.579	311	iyanello
Q8_0	0.75 GiB	804,753,664	10.801	—	Majipa
F16	1.12 GiB	1,197,629,632	16.074	310	Qwen
F16	1.12 GiB	1,197,634,272	16.074	311	iyanello

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	0.44 GiB	0.44 GiB	—	28 / 0 / 0
8,192	0.88 GiB	0.88 GiB	—	28 / 0 / 0
16,384	1.75 GiB	1.75 GiB	—	28 / 0 / 0
32,768	3.50 GiB	3.50 GiB	—	28 / 0 / 0
65,536	7.00 GiB	7.00 GiB	—	28 / 0 / 0
131,072	14.00 GiB	14.00 GiB	—	28 / 0 / 0

Compare with

same modality, comparable size

vs Llama-3.2-1B-Instruct vs ced-base vs gemma-3-1b-it vs Qwen3-1.7B vs embeddinggemma-300m vs Qwen3-0.6B 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 0.31 GiB. The real file is 0.37 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

1024

Vocab

151,936

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

false

Questions people ask

How much VRAM does Qwen3-0.6B-Base need?: Q4_K_M is exactly 396,476,256 bytes (0.37 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
How large is Qwen3-0.6B-Base's KV cache?: 3.50 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 Qwen3-0.6B-Base 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.