How much VRAM does Qwen3-Zero-Coder-Reasoning-V2-0.8B need?

I1-IQ1_S is exactly 255,783,232 bytes (0.24 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

How large is Qwen3-Zero-Coder-Reasoning-V2-0.8B's KV cache?

5.25 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.

DavidAU · text

Qwen3-Zero-Coder-Reasoning-V2-0.8B

DavidAU/Qwen3-Zero-Coder-Reasoning-V2-0.8B

Qwen3-Zero-Coder-Reasoning-V2-0.8B at I1-IQ1_S is exactly 255,783,232 bytes (0.24 GiB / 0.26 GB) — an effective 2.507 bits per weight, not the nominal 1. Its KV cache at 32K is 5.25 GiB.

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

Parameters

816M

Architecture

qwen3

42 layers

Context

40,960

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
I1-IQ1_S	0.24 GiB	255,783,232	2.507	—	mradermacher
I1-IQ1_M	0.25 GiB	267,800,896	2.625	—	mradermacher
I1-IQ2_XXS	0.27 GiB	287,830,336	2.821	—	mradermacher
I1-IQ2_XS	0.29 GiB	306,606,400	3.005	—	mradermacher
I1-IQ2_S	0.30 GiB	325,062,976	3.186	—	mradermacher
I1-IQ2_M	0.32 GiB	341,086,528	3.343	—	mradermacher
I1-Q2_K_S	0.33 GiB	354,426,176	3.474	—	mradermacher
I1-IQ3_XXS	0.34 GiB	362,074,432	3.549	—	mradermacher
I1-Q2_K	0.35 GiB	377,576,768	3.700	—	mradermacher
I1-IQ3_XS	0.37 GiB	402,496,832	3.945	—	mradermacher
I1-IQ3_S	0.39 GiB	417,832,256	4.095	—	mradermacher
I1-Q3_K_S	0.39 GiB	417,832,256	4.095	—	mradermacher
I1-IQ3_M	0.41 GiB	437,468,480	4.287	—	mradermacher
I1-Q3_K_M	0.42 GiB	453,975,360	4.449	—	mradermacher
I1-IQ4_XS	0.45 GiB	484,924,736	4.753	—	mradermacher
I1-Q3_K_L	0.45 GiB	485,956,928	4.763	—	mradermacher
I1-IQ4_NL	0.47 GiB	505,568,576	4.955	—	mradermacher
I1-Q4_0	0.47 GiB	506,551,616	4.965	—	mradermacher
I1-Q4_K_S	0.47 GiB	508,058,944	4.979	—	mradermacher
I1-Q4_K_M	0.49 GiB	528,276,800	5.177	—	mradermacher
I1-Q4_1	0.51 GiB	546,856,256	5.359	—	mradermacher
I1-Q5_K_S	0.55 GiB	588,143,936	5.764	—	mradermacher
I1-Q5_K_M	0.56 GiB	599,842,112	5.879	—	mradermacher
I1-Q6_K	0.63 GiB	675,880,256	6.624	—	mradermacher
Q8_03 shards	2.33 GiB	2,500,313,088	24.504	—	DavidAU
F163 shards	4.58 GiB	4,916,167,680	—	—	DavidAU

KV cache by context

computed per layer

Context	KV cache (f16)●	Flat formula	Overstated by	Full / windowed / recurrent
4,096	0.66 GiB	0.66 GiB	—	42 / 0 / 0
8,192	1.31 GiB	1.31 GiB	—	42 / 0 / 0
16,384	2.63 GiB	2.63 GiB	—	42 / 0 / 0
32,768	5.25 GiB	5.25 GiB	—	42 / 0 / 0
65,536	10.50 GiB	10.50 GiB	—	42 / 0 / 0
131,072	21.00 GiB	21.00 GiB	—	42 / 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 I1-IQ1_S at roughly 0.43 GiB. The real file is 0.24 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-Zero-Coder-Reasoning-V2-0.8B need?: I1-IQ1_S is exactly 255,783,232 bytes (0.24 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-Zero-Coder-Reasoning-V2-0.8B's KV cache?: 5.25 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-Zero-Coder-Reasoning-V2-0.8B 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.