Model comparison

canary-1b vs cohere-transcribe-03-2026

At Q4_K_M, canary-1b is the smaller download — 729,686,848 bytes against 1,558,162,944.

From the file· summed bytes, KV per layer

Side by side

canary-1bcohere-transcribe-03-2026
Parameters1.0B2.1B
Architecturecanarycohere_asr
Layers
Native context
Mixture of expertsnono
Quantizations published611
Smallest quantization0.68 GiB1.41 GiB
Q4_K_M0.68 GiB1.45 GiB
Licencecc-by-nc-4.0apache-2.0

KV cache by context

the term that decides long-context viability
Contextcanary-1bcohere-transcribe-03-2026Ratio
4,096
8,192
16,384
32,768
65,536
131,072