← All examples

ModelEthio-ASR Amharic (600M)17Ethio-ASR Multi (300M)19Ethio-ASR Multi (600M)13Ethio-ASR Multi (94M)19Meta MMS-1B19OmniASR-CTC (1B)19OmniASR-CTC (300M)19OmniASR-CTC (3B)19OmniASR-LLM (1B)15OmniASR-LLM (300M)10OmniASR-LLM (3B)10OpenAI Whisper-small19
Audio qualityClean Studio10Ambient Room Noise11Harsh Environment11Distant / Low-Volume Mic11Wideband Phone (VoLTE)10Randomized Phone Channel11Narrowband 2G + Packet Loss17

What did the model get wrong?

Ethio-ASR Amharic (600M) Narrowband 2G + Packet Loss   Example 1 of 20  ·  speaker: Female

🎧 Listen — Narrowband 2G + Packet Loss
Heavy background noise plus phone-call compression — worst-case stress test.
✅ What was actually said
ሁለት ዜሮ ሁለት አምስት አንድ ሁለት አንድ ሰባት ሁለት አራት ሰባት ስምንት ዜሮ ሰባት ዘጠኝ አራት
🤖 What this model heard
ኪን ሁት አም አንዱ መት ሁለት አራትራት ሰባትስምንትት ሰባትት ተን አራት
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
2
Correct
15
Wrong
2
Missed
0
Extra
11%
Words right
Out of 19 spoken words, this model got 2 right (11%). Errors: 15 words wrong; 2 missed.

Try switching Audio quality — see the same model degrade as the audio gets noisier.