โ† All examples

ModelEthio-ASR Amharic (600M)0Ethio-ASR Multi (300M)3Ethio-ASR Multi (600M)5Ethio-ASR Multi (94M)5Meta MMS-1B7OmniASR-CTC (1B)0OmniASR-CTC (300M)1OmniASR-CTC (3B)0OmniASR-LLM (1B)0OmniASR-LLM (300M)1OmniASR-LLM (3B)1OpenAI Whisper-small12
Audio qualityClean Studio1Ambient Room Noise3Harsh Environment2Distant / Low-Volume Mic1Wideband Phone (VoLTE)4Randomized Phone Channel8Narrowband 2G + Packet Loss9

What did the model get wrong?

OmniASR-CTC (300M) Clean Studio   Example 14 of 20  ยท  speaker: Male

๐ŸŽง Listen โ€” Clean Studio
Clear, studio-quality recording. Best-case audio.
โœ… What was actually said
แŠฅแˆฝ แ‰ชแ‹ตแ‹ฎ แ‹ˆแ‹ญแˆ แˆ˜แˆ˜แˆชแ‹ซ แŠ แˆˆ แŠฅแŠ•แ‹ฐแ‰ต แŠฅแŠ•แ‹ฐแˆแŒ แ‰€แˆ
๐Ÿค– What this model heard
แŠฅแˆฝ แ‰ชแ‹ตแ‹ฎ แ‹ˆแ‹ญแˆ แˆ˜แˆ˜แˆช แŠ แˆˆ แŠฅแŠ•แ‹ฐแ‰ต แŠฅแŠ•แ‹ฐแˆแŒ แ‰€แˆ
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
6
Correct
1
Wrong
0
Missed
0
Extra
86%
Words right
Out of 7 spoken words, this model got 6 right (86%). Errors: 1 word wrong.

Try switching Audio quality โ€” see the same model degrade as the audio gets noisier.