Multilingual
SUTRA-PRO
and SUTRA-LIGHT
are multilingual models, represents a broad array of multilingual AI capabilities. Developed with a rich diversity of instructions from both proprietary and leading open-access datasets, our models excel across a spectrum of multilingual benchmarks, offering unparalleled proficiency in over 50 languages including the nuanced spectrums of Latin, Indic, and Far Eastern languages.
SUTRA models are powerful tools capable of deriving insights from text data across languages. We have developed these models specifically for use in conversations, search, summarization, and numerous other tasks.
Conversation
SUTRA-LIGHT
facilitates fluid, natural conversations across a vast array of languages, breaking down barriers and fostering global connections.
Question: “What are the main tourist attractions in Rome?”
SUTRA-LIGHT
English Response
SUTRA-LIGHT
Hindi Response
SUTRA-LIGHT
Gujarati Response
SUTRA-LIGHT
Tamil Response
SUTRA-LIGHT
Korean Response
Instruction following with Precision
From cooking recipes to complex data analysis, SUTRA-LIGHT
understands and executes detailed instructions across languages with impressive accuracy. This makes them incredibly useful tools for global communication and content creation. To illustrate, let’s see how an LLM can handle a simple task across different languages: providing the current time in a specified city.
Instruction in English: “List the steps to make a simple avocado toast”
Instruction in Hindi: “एवोकाडो टोस्ट बनाने की विधि की सूची”
Instruction in Gujarati: “એવોકાડો ટોસ્ટ બનાવવાની યાદી પદ્ધતિ”
Instruction in Tamil: “அவகேடோ டோஸ்ட் செய்யும் முறை பட்டியல்”
Instruction in Korean: “간단한 아보카도 토스트 만드는 방법을 나열하시오”
Translation
With SUTRA-LIGHT
, experience translation that captures not just the words, but the essence of your message, maintaining cultural nuances and contextual accuracy - supporting not just native script but also colloquial communication like Hinglish
SUTRA-LIGHT
Output for the question: “What is the fastest way to travel from Mumbai to Delhi?”
English:
Hinglish:
Gujarati:
Tamil: