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

Q4_K_M is exactly 589,560,480 bytes (0.55 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

Qwen · audio asr

Qwen3-ASR-0.6B

Qwen/Qwen3-ASR-0.6B

Qwen3-ASR-0.6B at Q4_K_M is exactly 589,560,480 bytes (0.55 GiB / 0.59 GB) — an effective 5.028 bits per weight, not the nominal 4.

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

Parameters

938M

Architecture

qwen3_asr

null layers

Context

—

native (config.json)

License

apache-2.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_K_M	0.55 GiB	589,560,480	5.028	613	handy-computer
Q4_K	0.59 GiB	631,026,336	5.382	—	cstr
Q5_K_M	0.60 GiB	645,356,192	5.504	613	handy-computer
Q6_K	0.64 GiB	690,417,824	5.888	613	handy-computer
Q8_0	0.75 GiB	804,749,248	6.864	311	ggml-org
Q8_0	0.79 GiB	850,423,456	7.253	—	handy-computer
Q8_0	0.94 GiB	1,006,809,760	8.587	—	cstr
BF16	1.41 GiB	1,509,343,168	12.873	—	ggml-org
BF16	1.46 GiB	1,571,490,016	13.403	—	handy-computer
F16	1.47 GiB	1,579,793,056	13.474	613	handy-computer

Measured

published by a third party, attributed below

Metric	Value	What it means
rtf	438.7
RTFx	438.7	Higher is better — audio seconds processed per second of compute.
Word error rate	9.47%	Lower is better — the share of words transcribed incorrectly.
Word error rate	6.80%	Lower is better — the share of words transcribed incorrectly.
Word error rate	10.72%	Lower is better — the share of words transcribed incorrectly.
Word error rate	7.62%	Lower is better — the share of words transcribed incorrectly.
Word error rate	1.69%	Lower is better — the share of words transcribed incorrectly.
Word error rate	3.97%	Lower is better — the share of words transcribed incorrectly.
Word error rate	2.74%	Lower is better — the share of words transcribed incorrectly.
Word error rate	3.07%	Lower is better — the share of words transcribed incorrectly.
Word error rate	10.57%	Lower is better — the share of words transcribed incorrectly.
Word error rate	7.65%	Lower is better — the share of words transcribed incorrectly.
Word error rate	5.61%	Lower is better — the share of words transcribed incorrectly.

Benchmarked· by open-asr-leaderboard-english-short-latest

RTFx measured by the Open ASR Leaderboard on a single datacenter GPU at a large batch size. It ranks models against each other; it says nothing about throughput on consumer hardware. We reproduce these figures with attribution; they are not ours and we have not verified the runs. Source: open-asr-leaderboard-english-short-latest.

Compare with

same modality, comparable size

vs nemotron-3.5-asr-streaming-0.6b vs parakeet-unified-en-0.6b vs parakeet-tdt-0.6b-v3 vs whisper-medium vs whisper-large-v3 vs canary-180m-flash 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 0.49 GiB. The real file is 0.55 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

—

Hidden size

—

Vocab

—

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does Qwen3-ASR-0.6B need?: Q4_K_M is exactly 589,560,480 bytes (0.55 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of Qwen3-ASR-0.6B 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.