Side by side

See the difference.

Scores, capabilities and costs for the models you’re considering.

Comparison of HojoAI/Hojo-ASR-V1
1 modelHojoAIHojoAI/Hojo-ASR-V1Remove ×
What’s different5.18B parameters · 2 languages · Encoder: Qwen3-Omni encoder · Decoder: Qwen3-4B
DateOct 17, 2026Discovered
Listed rateSee provider quotes
Provider endpoints0 endpoints →
AMI-Cleaned WER ↗English short-form; normalization unversioned7.54% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
Earnings22-Cleaned-AA-chunked WER ↗English short-form; normalization unversioned6.97% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
Gigaspeech-Cleaned WER ↗English short-form; normalization unversioned6.14% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
LS Clean WER ↗English short-form; normalization unversioned1.3% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
LS Other WER ↗English short-form; normalization unversioned3.24% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
SPGISpeech WER ↗English short-form; normalization unversioned1.74% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
URGENT2024 WER ↗English short-form; normalization unversioned12.33% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
Voice Arena Monsoon WER ↗English short-form; normalization unversioned4.63% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
Voxpopuli-AA-Cleaned WER ↗English short-form; normalization unversioned3.11% WER ↗HojoAI/Hojo-ASR-V1Leaderboard entry
Context window—
Max output—
Parameters5.18B
Audio inputSupported
TranscriptionSupported

Add another model

Search all models →