Source data
BHASHINI’s language data, models and service capabilities
Sources checked
- 01Parallel sentence pairs
246 million pairs · Text aligned across languages for translation work Source 1
- 02Monolingual text
3.7 million entries · Text resources in a single language Source 1
- 03AI language model repository
More than 360 models · Models supporting specialised language tasks Source 1
- 04Language services
More than 22 specialised services · Functions include recognition, translation and speech generation Source 1
- 05Service coverage, March 2026
36 text languages and 23 voice languages · Coverage varies by modality Source 2
- 06Shrutlekh
22 Indian languages · Live transcription and translation tool Source 3
The flow shows where data, models and applications fit. Individual applications can use different paths.
BHASHINI’s language data, models and service capabilities
Published figures describe separate stages or capabilities and should not be added. The speech and text language counts refer to the March programme update. Corpus and model counts come from an August parliamentary reply.
| Source | |||
|---|---|---|---|
| Parallel sentence pairs | 246 million pairs | Text aligned across languages for translation work | Source 1 |
| Monolingual text | 3.7 million entries | Text resources in a single language | Source 1 |
| AI language model repository | More than 360 models | Models supporting specialised language tasks | Source 1 |
| Language services | More than 22 specialised services | Functions include recognition, translation and speech generation | Source 1 |
| Service coverage, March 2026 | 36 text languages and 23 voice languages | Coverage varies by modality | Source 2 |
| Shrutlekh | 22 Indian languages | Live transcription and translation tool | Source 3 |