हिंदी में पढ़ें —JantaScope हिंदी
AI NEWS

Sarvam Cuts Vision AI Model API Prices by 67% as Document Digitisation Crosses 35 Million Pages

Sarvam AI has reduced the price of its Vision API by 67% after crossing 35 million digitised pages through its platform. Indian AI startup Sarvam cuts Vision API costs after massive adoption and document digitisation growth.

Sarvam Cuts Vision AI Model API Prices by 67% as Document Digitisation Crosses 35 Million Pages

By Jeet Nirmal

Source: Janta Scope

Bengaluru-based AI startup Sarvam AI has announced a major 67% reduction in the pricing of its Sarvam Vision API after the platform crossed 35 million digitised pages, marking a significant milestone in India's growing AI ecosystem.

The company reduced the cost of its Vision API from ₹1.5 per page to ₹0.5 per page, saying improvements in infrastructure efficiency and large-scale deployment helped lower operational costs. Sarvam stated that the savings are now being passed directly to developers and enterprise customers.

Launched in February 2026, Sarvam Vision is a vision-language AI model built for document intelligence tasks such as OCR, document digitisation, text extraction, and converting physical records into searchable digital formats. Since launch, the platform has seen rapid adoption among businesses, developers, and enterprise partners

According to the company, more than 35 million pages have already been processed through the platform. Sarvam says it achieved the price reduction through improvements in AI serving infrastructure, smarter batching systems, optimized inference performance, and better hardware utilization.

The development comes as Sarvam continues positioning itself as one of India's leading AI startups. Earlier this year, the company claimed that Sarvam Vision outperformed several global AI models in OCR-related benchmarks, particularly for Indian languages and document processing tasks.

Related

More stories

Fake ChatGPT, Claude and Gemini Apps Are Spreading Malware — Users Warned About AI Impersonation Threats

Cybercriminals are increasingly disguising malicious software as popular AI services such as ChatGPT, Claude and Gemini. Kaspersky says its security systems detected more than 92,000 attacks involving malware or potentially unwanted applications masquerading as AI services between January and early May 2026, highlighting the growing security risks surrounding the rapid adoption of artificial intelligence.

AI NEWS

Fake ChatGPT, Claude and Gemini Apps Are Spreading Malware — Users Warned About AI Impersonation Threats

IIT Madras Builds AI Platform With 185,000 Alloy Records, Targets Faster Discovery of Greener Materials

Researchers at IIT Madras have developed an artificial intelligence-driven platform designed to accelerate the discovery of sustainable, high-performance metallic alloys. The project has produced two databases containing more than 185,000 structured records, extracted from over 10,000 scientific papers, with potential applications spanning electric vehicles, aerospace, renewable energy and marine infrastructure.

AI NEWS

IIT Madras Builds AI Platform With 185,000 Alloy Records, Targets Faster Discovery of Greener Materials

Meta Hires OpenAI Veteran Luke Metz for Superintelligence Labs in Major AI Talent Push

Meta has hired prominent artificial intelligence researcher Luke Metz to join its Superintelligence Labs. Metz, who previously worked at OpenAI and Mira Murati’s Thinking Machines Lab, is expected to report to Alexandr Wang as Meta continues strengthening its team in the intensifying race to develop advanced AI systems.

AI NEWS

Meta Hires OpenAI Veteran Luke Metz for Superintelligence Labs in Major AI Talent Push

Nvidia Is No Longer Just an AI Chip Giant — Its Push Into AI Models Is Getting Much Bigger

Nvidia is expanding beyond the hardware that powered the generative-AI boom and strengthening its position in AI models, particularly through its Nemotron family and a major deal involving AI startup Poolside. The strategy could give Nvidia greater influence across the entire AI technology stack—from computing infrastructure to the models and agents running on top of it.

AI NEWS

Nvidia Is No Longer Just an AI Chip Giant — Its Push Into AI Models Is Getting Much Bigger

Anthropic’s $2 Trillion IPO Dream Could Rewrite Wall Street Records — Here’s Why Investors Are Watching

Artificial intelligence company Anthropic is moving closer to a potential stock-market debut that could become one of the largest IPOs ever. Investors are reportedly discussing a valuation of around $2 trillion or more, highlighting extraordinary expectations surrounding the company behind the Claude AI models.

AI NEWS

Anthropic’s $2 Trillion IPO Dream Could Rewrite Wall Street Records — Here’s Why Investors Are Watching

Meta’s AI Talent War Heats Up Again as OpenAI Veteran Luke Metz Makes the Switch

The battle for the world’s most sought-after artificial intelligence researchers is intensifying again. Meta has hired veteran AI researcher Luke Metz, adding another former OpenAI talent to its expanding Superintelligence Labs operation. The move highlights how competition between leading AI companies is increasingly being fought not only through bigger models and computing infrastructure, but also through the recruitment of a relatively small group of researchers with experience building front

AI NEWS

Meta’s AI Talent War Heats Up Again as OpenAI Veteran Luke Metz Makes the Switch