Post Thumbnail

Supercomputer and 12 billion parameters: Krutrim opens a new chapter in Indian AI

Ola company founder and billionaire Bhavish Aggarwal announced investing $230 million in his AI startup Krutrim, intending to make India a leader in the global artificial intelligence race. According to informed sources at TechCrunch, the funding will be primarily through Aggarwal’s family office, with total investments planned to reach $1.15 billion by next year.

Krutrim marked an important event by opening access to its AI models and announcing the construction of India’s largest supercomputer in partnership with Nvidia. The company introduced Krutrim-2 – a language model with 12 billion parameters, specializing in processing Indian languages.

The test results are impressive: in text sentiment analysis, the model showed a result of 0.95 compared to 0.70 from competitors, and in code generation tasks achieved 80% success. Technical capabilities include a context window of 128,000 tokens, allowing for processing long texts and conducting complex dialogues. In grammar correction tests, the model reached 0.98, and in multi-stage conversations – 0.91.

“We are still far from global standards, but have achieved good progress in a year,” noted Aggarwal. “By opening access to our models, we hope for collaboration from the entire Indian AI community to create a world-class ecosystem.”

Krutrim has also developed its own BharatBench evaluation system for testing AI models’ effectiveness in working with Indian languages, filling a gap in existing evaluation systems that primarily focus on English and Chinese languages.

The initiative emerged against the backdrop of India’s desire to strengthen its position in artificial intelligence, where American and Chinese companies dominate. Recently, India welcomed the progress of Chinese company DeepSeek and announced the placement of its language models on domestic servers. Krutrim’s cloud division has already begun providing access to DeepSeek on Indian servers.

Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

Latest News

AI fraudster posed as Keanu Reeves for 2.5 years and stole $160,000

Dianne Ringstaff became a victim of sophisticated fraud when she was playing a mobile game and received a message followed by a video call. Artificial intelligence technologies were so advanced that the woman was absolutely certain — the real Keanu Reeves was calling her.

70-kilogram humanoid flies on turbines

Italian engineers accomplished the incredible. The metallic flying humanoid iRonCub3 with human proportions weighing 70 kg flies! 4 powerful turbines lifted the humanoid to a height of 50 cm, demonstrating technology that previously existed only in science fiction.

MIT and Microsoft exposed GPT-3.5's lies

A team of scientists from MIT and Microsoft developed a methodology that allows looking behind the scenes of language models' thinking. And understanding when they lie to us. The research reveals disturbing cases of systematic mismatch between the real reasons for models' decisions and their verbal explanations.

OpenAI lures Microsoft clients with discounts

OpenAI company began providing significant discounts on ChatGPT corporate subscriptions — from 10 to 20%! But discounts are available with additional investments in other OpenAI products, including Deep Research, Codex and increased API expenses. And this unprecedented move causes serious concern at Microsoft.

GigaChat lost to Claude and Gemini in Russian language on MERA benchmark

Testing GigaChat reveals the harsh truth about Russia's place in the global artificial intelligence race. Recent tests on the MERA benchmark showed results that force serious reflection. The Russian model, created specifically for working with Russian language, unexpectedly lost to foreign competitors in its own "native element".