Model comparison

canary-180m-flash vs whisper-small

At Q4_K_M, canary-180m-flash is the smaller download — 139,223,744 bytes against 171,630,656.

From the file· summed bytes, KV per layer

Side by side

canary-180m-flashwhisper-small
Parameters189M242M
Architecturecanarywhisper
Layers12
Native context
Mixture of expertsnono
Quantizations published619
Smallest quantization0.13 GiB0.08 GiB
Q4_K_M0.13 GiB0.16 GiB
Licencecc-by-4.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextcanary-180m-flashwhisper-smallRatio
4,096
8,192
16,384
32,768
65,536
131,072