Open multilingual ASR dark horse: Hojo-ASR-Multi-V1 hits 3.54% WER, ranks #7 globally
rohanpaul_ai · x · 2026-09-04
HojoAI's Hojo-ASR-Multi-V1 debuts at #7 globally and #1 among open multilingual models on the Open ASR Leaderboard, with an average WER of 3.54% across five languages, rivaling leading proprietary speech recognition systems. Rohan Paul frames it through an AI-hardware lens: smartphones' biggest limitation is demanding attention, and specialized devices that minimize interaction for one job are newly economically viable.
- Open weights, public eval files, reproducible results
- Supports major languages including German, French, Spanish, Japanese, Arabic, Korean and Russian
Related event: Hojo-ASR Tops Open Multilingual ASR Leaderboard with 3.54% WER(2 posts)→
More from Embodied
- Stanford's T2-z lets robots probe hidden physics with a poke, lifting success from 34% to 73% — stanfordnlp · 2026-10-21
- Snap Hires Robert Downey Jr. as Chief Customer Experience Officer to Sell Specs AR Glasses Against Meta — adariostrange · 2026-10-21
- Fei-Fei Li team unveils FACT, a force-aware foundation model for robot dexterity — stanfordnlp · 2026-10-21
- Upcoming paper claims big models can handle latency-critical real-time robotics — ManlingLi_ · 2026-10-21
- New project at the intersection of agents, embodied AI, and games gets a nod — ZeeshanZiaML · 2026-10-21
- NVIDIA open-sources Flora, a robotics data engine 125x cheaper than GPT-6 Astra — DrJimFan · 2026-10-21