Alternatives to canary-qwen-2.5b

canary-qwen-2.5b alternatives

canary-qwen-2.5b's smallest published quantization is 1.62 GiB. The models below do the same job with a different trade — less memory, comparable size, or a more permissive licence.

From the file· sizes from summed file bytes

Meaningfully smaller

under 70% of its smallest quantization
ModelParamsSmallestQuantsLicence
nemotron-3.5-asr-streaming-0.6b638M0.38 GiB8other
parakeet-unified-en-0.6b618M0.44 GiB6cc-by-4.0
parakeet-tdt-0.6b-v3627M0.39 GiB10cc-by-4.0
whisper-medium764M0.25 GiB19apache-2.0
whisper-large-v31.5B0.49 GiB18apache-2.0
canary-180m-flash189M0.13 GiB6cc-by-4.0
whisper-large-v3-turbo809M0.27 GiB31apache-2.0
Qwen3-ASR-0.6B938M0.55 GiB10apache-2.0

Comparable in size

within ±40%, so a like-for-like swap
ModelParamsSmallestQuantsLicence
cohere-transcribe-03-20262.1B1.41 GiB11apache-2.0
Qwen3-ASR-1.7B2.3B1.23 GiB13apache-2.0
granite-speech-4.1-2b-nar2.3B1.45 GiB8apache-2.0
Voxtral-Mini-3B-25074.7B1.45 GiB31apache-2.0
granite-4.0-1b-speech2.3B1.49 GiB6apache-2.0

More permissively licensed

licences that allow commercial use
ModelParamsSmallestQuantsLicence
parakeet-unified-en-0.6b618M0.44 GiB6cc-by-4.0
cohere-transcribe-03-20262.1B1.41 GiB11apache-2.0
parakeet-tdt-0.6b-v3627M0.39 GiB10cc-by-4.0
whisper-medium764M0.25 GiB19apache-2.0
Voxtral-Mini-4B-Realtime-26024.4B2.35 GiB8apache-2.0
whisper-large-v31.5B0.49 GiB18apache-2.0
canary-180m-flash189M0.13 GiB6cc-by-4.0
whisper-large-v3-turbo809M0.27 GiB31apache-2.0