Model reference · open weights

indic-conformer-multilingual

indic-conformer-multilingual is an open-weight audio or speech model from ai4bharat.

• IndicConformer-600M-Multi is an automatic speech recognition model developed by ai4bharat for transcribing audio in 22 official Indian languages.
• The model utilizes a multilingual Conformer-based hybrid architecture with CTC and RNNT decoding strategies and contains 600 million parameters.
• It is released under the MIT license.

Summary of the ai4bharat/indic-conformer-600m-multilingual model card, 2026-10-01

Hugging Face ↗