โ† All examples

ModelEthio-ASR Amharic (600M)0Ethio-ASR Multi (300M)2Ethio-ASR Multi (600M)1Ethio-ASR Multi (94M)4Meta MMS-1B8OmniASR-CTC (1B)1OmniASR-CTC (300M)3OmniASR-CTC (3B)1OmniASR-LLM (1B)3OmniASR-LLM (300M)4OmniASR-LLM (3B)5OpenAI Whisper-small21
Audio qualityClean Studio8Ambient Room Noise13Harsh Environment21Distant / Low-Volume Mic8Wideband Phone (VoLTE)8Randomized Phone Channel8Narrowband 2G + Packet Loss9

What did the model get wrong?

OpenAI Whisper-small Harsh Environment   Example 15 of 20  ยท  speaker: Female

๐ŸŽง Listen โ€” Harsh Environment
Very heavy background noise โ€” much harder than typical audio.
โœ… What was actually said
แŠฅแˆฝ แŠ แˆแŠ• แŠฅแŠ•แ‹ฐแ‰ต แŠแ‹ แ‹จแˆแŒ แ‰€แˆ˜แ‹ แŠ แ‹ตแˆต แ‰แŒฅแˆญ แŠ แˆแ‹ฐแˆจแˆฐแŠแˆ
๐Ÿค– What this model heard
y shi ahun ahun ahon ahon ahon and it is no mette kamo ndi tte kamo adiskutor adiskutor aldera aldera raseinim
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
0
Correct
8
Wrong
0
Missed
13
Extra
0%
Words right
Out of 8 spoken words, this model got 0 right (0%). Errors: 8 words wrong; 13 extra.

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