Model comparison

parakeet-unified-en-0.6b vs whisper-large-v3

At Q4_K_M, parakeet-unified-en-0.6b is the smaller download — 477,274,496 bytes against 997,303,008.

From the file· summed bytes, KV per layer

Side by side

parakeet-unified-en-0.6bwhisper-large-v3
Parameters618M1.5B
Architectureparakeetwhisper
Layers32
Native context
Mixture of expertsnono
Quantizations published618
Smallest quantization0.44 GiB0.49 GiB
Q4_K_M0.44 GiB0.93 GiB
Licencecc-by-4.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextparakeet-unified-en-0.6bwhisper-large-v3Ratio
4,096
8,192
16,384
32,768
65,536
131,072