โ† All examples

ModelEthio-ASR Amharic (600M)5Ethio-ASR Multi (300M)6Ethio-ASR Multi (600M)4Ethio-ASR Multi (94M)3Meta MMS-1B12OmniASR-CTC (1B)9OmniASR-CTC (300M)9OmniASR-CTC (3B)3OmniASR-LLM (1B)0OmniASR-LLM (300M)3OmniASR-LLM (3B)0OpenAI Whisper-small9
Audio qualityClean Studio3Ambient Room Noise0Harsh Environment0Distant / Low-Volume Mic3Wideband Phone (VoLTE)0Randomized Phone Channel1Narrowband 2G + Packet Loss11

What did the model get wrong?

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

๐ŸŽง Listen โ€” Clean Studio
Clear, studio-quality recording. Best-case audio.
โœ… What was actually said
แ‰แŒฅแˆฉ แ‹œแˆฎ แ‹˜แŒ แŠ แˆแˆˆแ‰ต แ‹œแˆฎ แŠ แˆแˆตแ‰ต แˆตแ‹ตแˆตแ‰ต แˆฐแ‰ฃแ‰ต แˆตแˆแŠ•แ‰ต แ‹˜แŒ แŠ แ‹œแˆฎ แŠแ‹
๐Ÿค– What this model heard
แ‰แŒฅแˆฉ แ‹œแˆฎ แ‹˜แŒ แŠ แˆแˆˆแ‰ต แ‹œแˆฎ แˆƒแˆแˆณ แˆฐแ‰ฃแ‰ต แˆฐแˆ›แŠ•แ‹ซ แ‹˜แŒ แŠ แ‹œแˆฎ แŠแ‹
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
9
Correct
2
Wrong
1
Missed
0
Extra
75%
Words right
Out of 12 spoken words, this model got 9 right (75%). Errors: 2 words wrong; 1 missed.

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