1 story in the last 7 days
The latest prisma v2.5 news, distilled by AI into sharp ~100-word summaries. ByteBrief tracks prisma v2.5 across dozens of tech sources and brings you only what matters, updated hourly. Tap any story for the full brief, or open the original source.
Gnani.ai released Prisma v2.5, a speech-to-text model supporting 12 languages and trained on 14 million hours of proprietary Indic speech data. The model handles dialect variation, background noise, and code-switching, and is available via APIs. It targets accuracy improvements for short utterances, numbers, and named entities in banking, insurance, and healthcare.
Summaries by ByteBrief