Chariot Launches Speech Recognition Model for 10 Indian Languages
Chariot's speech-to-text model converts spoken words to text in real-time across Hindi, Marathi, Bengali, Tamil, Telugu, Kannada, Malayalam, Gujarati, Punjabi and Odia. Testing on 79 public datasets with 153,000 audio samples including conversations, calls and noisy environments showed the model achieved 26% fewer errors than Google Gemini 3.5 Transcribe and 34% fewer than its Live variant. Against real-time systems, it showed 17% fewer errors than Sarvam's competing model and 57% fewer than ElevenLabs Scribe v2.
Key points
- Supports 10 Indian languages plus Indian English in real-time speech-to-text
- Tested on 153,000 audio samples covering diverse conditions
- Claims 26-34% lower error rates versus Google Gemini and 17-57% versus other streaming systems
Summary and key points written by AI from the headline and a short excerpt, so they can be incomplete or wrong. Analytics India Magazine did the reporting; read the original for the full story.
Read the full story at Analytics India Magazine
