Side by side

See the difference.

Scores, capabilities and costs for the models you’re considering.

Comparison of Edge0/ARK-ASR-0.6B
1 modelEdge0Edge0/ARK-ASR-0.6BRemove ×
What’s different1.15B parameters · 19 languages · Encoder: Whisper · Decoder: Qwen2
DateOct 17, 2026Discovered
Listed rateSee provider quotes
Provider endpoints0 endpoints →
AMI-Cleaned WER ↗English short-form; normalization unversioned8.61% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned6.75% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned7.37% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
LS Clean WER ↗English short-form; normalization unversioned1.48% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
LS Other WER ↗English short-form; normalization unversioned3.44% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
SPGISpeech WER ↗English short-form; normalization unversioned2.63% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
URGENT2024 WER ↗English short-form; normalization unversioned12.49% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
Voice Arena Monsoon WER ↗English short-form; normalization unversioned3.43% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned2.76% WER ↗Edge0/ARK-ASR-0.6BLeaderboard entry
Context window—
Max output—
Parameters1.15B
Audio inputSupported
TranscriptionSupported

Add another model

Search all models →