How much VRAM does parakeet-ctc-0.6b need?

Q4_K_M is exactly 469,302,464 bytes (0.44 GiB) in weights. Add the KV cache, which depends on context length, plus roughly half a gigabyte of runtime overhead.

nvidia · audio asr

parakeet-ctc-0.6b

nvidia/parakeet-ctc-0.6b

parakeet-ctc-0.6b at Q4_K_M is exactly 469,302,464 bytes (0.44 GiB / 0.47 GB) — an effective 6.166 bits per weight, not the nominal 4.

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

Parameters

609M

Architecture

parakeet

null layers

Context

—

native (config.json)

License

cc-by-4.0

Shipped quantizations

● exact bytes, summed from published files

Quant	Size●	Exact bytes●	Effective bpw●	Tensors●	Publisher
Q4_K_M	0.44 GiB	469,302,464	6.166	950	handy-computer
Q5_K_M	0.50 GiB	532,544,704	6.997	950	handy-computer
Q6_K	0.55 GiB	593,644,736	7.800	950	handy-computer
Q8_0	0.67 GiB	722,271,424	9.490	950	handy-computer
F16	1.14 GiB	1,220,181,184	16.033	950	handy-computer
F32	2.27 GiB	2,435,482,816	32.001	—	handy-computer
Q4_K11 shards	7.34 GiB	7,883,768,736	—	—	mudler
Q5_K11 shards	8.11 GiB	8,704,557,984	—	—	mudler
Q6_K11 shards	8.92 GiB	9,576,646,560	—	—	mudler
Q8_011 shards	10.39 GiB	11,156,097,024	—	—	mudler
F1611 shards	16.13 GiB	17,315,472,384	—	—	mudler

Measured

published by a third party, attributed below

Metric	Value	What it means
rtf	5883.9
RTFx	5883.9	Higher is better — audio seconds processed per second of compute.
Word error rate	3.41%	Lower is better — the share of words transcribed incorrectly.
Word error rate	3.47%	Lower is better — the share of words transcribed incorrectly.
Word error rate	3.19%	Lower is better — the share of words transcribed incorrectly.
Word error rate	15.56%	Lower is better — the share of words transcribed incorrectly.
Word error rate	6.77%	Lower is better — the share of words transcribed incorrectly.
Word error rate	9.03%	Lower is better — the share of words transcribed incorrectly.
Word error rate	13.10%	Lower is better — the share of words transcribed incorrectly.
Word error rate	13.47%	Lower is better — the share of words transcribed incorrectly.
Word error rate	8.90%	Lower is better — the share of words transcribed incorrectly.
Word error rate	1.57%	Lower is better — the share of words transcribed incorrectly.
Word error rate	6.73%	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 0.32 GiB. The real file is 0.44 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

1,025

Sliding window

none

SWA period

—

MLA

Experts

—

Experts per token

—

use_sliding_window

—

Questions people ask

How much VRAM does parakeet-ctc-0.6b need?: Q4_K_M is exactly 469,302,464 bytes (0.44 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of parakeet-ctc-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.