โ† All examples

ModelEthio-ASR Amharic (600M)0Ethio-ASR Multi (300M)4Ethio-ASR Multi (600M)2Ethio-ASR Multi (94M)2Meta MMS-1B7OmniASR-CTC (1B)1OmniASR-CTC (300M)1OmniASR-CTC (3B)3OmniASR-LLM (1B)3OmniASR-LLM (300M)3OmniASR-LLM (3B)3OpenAI Whisper-small13
Audio qualityClean Studio1Ambient Room Noise3Harsh Environment3Distant / Low-Volume Mic1Wideband Phone (VoLTE)3Randomized Phone Channel9Narrowband 2G + Packet Loss7

What did the model get wrong?

OmniASR-LLM (3B) Ambient Room Noise   Example 10 of 20  ยท  speaker: Male

๐ŸŽง Listen โ€” Ambient Room Noise
Realistic everyday background noise โ€” like a normal room or cafรฉ.
โœ… What was actually said
แˆตแˆˆ แŒฅแ‰…แˆŽแ‰น แ‹แˆญแ‹แˆญ แˆ˜แˆจแŒƒ แ‹จแ‰ต แŠแ‹ แ‹จแˆ›แŒˆแŠ˜แ‹
๐Ÿค– What this model heard
แˆตแˆˆ แŒฅแ‰…แˆŽแ‰น แ‹แˆญแ‹แˆญ แˆ˜แˆจแŒƒ แ‹จ แ‰ต แŠแ‹ แˆ›แŒˆแŠ˜แ‹
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
5
Correct
2
Wrong
0
Missed
1
Extra
71%
Words right
Out of 7 spoken words, this model got 5 right (71%). Errors: 2 words wrong; 1 extra.

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