BodhanAI and AI4Bharat launch Indic-Speech: one 4B model, 23 languages, code-mixed transcription
sumanthd17 · x · 2026-09-08
BodhanAI and AI4Bharat announced Indic-Speech: a single 4B-parameter model covering 23 languages for transcription, translation, recognition, and synthesis.
The standout feature is code-switching: it can transcribe sentences naturally mixing English and Tamil (e.g. "I'm fluent in English and Tamil. அதனால நான் இப்போ Tamil-இல் பேசுறதும்"). It also supports two-speaker podcasts and expressive narration. Goes live on Bodhan AI on 05.09.2026, coming soon to muesli.
More from Models
- inclusionAI's Ling-3.0-flash-Fin, a Finance-Tuned MoE Model, Trends on Hugging Face — inclusionAI · 2026-09-08
- Chess AI Astra beats Wally with 6 extra minutes; author estimates it's now well above 1800 ELO — MikePFrank · 2026-09-08
- Heavy user says AI benchmarks are now 'harmful to trust', not just misleading — aziham · 2026-09-08
- GLM-5.3 ships on same base as 5.2: post-training alone lifts Terminal Bench 4.6→28.3 — burny_tech · 2026-09-08
- A provably optimal strategy for juggling Claude and Codex subscription quotas — DimitrisPapail · 2026-09-08
- Gemini 4.0 Pro could match Astra at 50% of its price — bindureddy · 2026-09-08