How do I use Nvidia backs Sarvam AI models while Google uses Gemini AI?

Nvidia backs Sarvam AI models while Google uses Gemini AI can be accessed through the provided link. Follow the instructions on the tool's website to get started. Most AI tools offer intuitive interfaces designed for easy use.

Nvidia backs Sarvam AI models while Google uses Gemini AI

Q: What does Nvidia backs Sarvam AI models while Google uses Gemini AI cost?

The pricing for Nvidia backs Sarvam AI models while Google uses Gemini AI is: Check the official website. Visit the official website for the most up-to-date pricing information and available plans.

Q: What can I use Nvidia backs Sarvam AI models while Google uses Gemini AI for?

Nvidia backs Sarvam AI models while Google uses Gemini AI is designed for analysis applications. It helps users accomplish tasks related to these areas efficiently and effectively.

Indian startup Sarvam AI recently unveiled its new 30-billion and 105-billion parameter AI models on February 18, 2026, at an AI summit in New Delhi. These models, built from scratch in India with support from Nvidia and others, are designed for India's diverse languages, including Hinglish. Sarvam AI, which has secured over $50 million in funding, also introduced text-to-speech, speech-to-text, and vision models, aiming to foster a "sovereign AI" ecosystem in the country.

The broader AI landscape sees varied developments, from specialized tools to new social dynamics. BLACKBOX.AI, for instance, offers an AI coding assistant that translates natural language into functional code, integrating with VS Code and web browsers to speed up software development. Meanwhile, AI toy company Miko responded to public pressure by adding an on-off switch for AI features in its Miko 3 and Miko Mini robots on February 17, 2026, following privacy concerns. These Miko toys utilize Google Cloud and Gemini AI models.

Concerns about AI's presence online are also growing, with a CNET survey revealing that less than half of adults believe they can identify AI-generated content. In response, Google's President of Global Affairs, Kent Walker, promoted an AI-first security strategy at the Munich Security Conference 2026, emphasizing AI as a defense against cyber threats. A novel platform called RentAHuman, launched on February 1, 2026, allows AI agents like Claude and GPT-4 to hire humans for real-world tasks, with over half a million people already signed up.

AI is also making strides in scientific research and career development. University of Missouri researchers created PSBench, the world's largest database of 1.4 million verified protein models, which can accelerate drug development for diseases like Alzheimer's and cancer. For individuals, online schools like Curious Refuge are helping professionals, such as laid-off VFX artist Michael Eng, transition into AI-powered filmmaking, training thousands in new machine learning skills.

Further showcasing AI's evolving role, Moltbook, a new social network for AI personal assistants, launched on February 12, 2026. This platform enables bots to communicate, develop unique personas, and even engage in an "agent economy" where they trade code and coordinate projects. This highlights the emerging complexity and social interactions within AI systems.

Key Takeaways

Sarvam AI launched new 30B and 105B parameter models on February 18, 2026, built in India with Nvidia support, and secured over $50 million in funding.
Sarvam AI's models are designed for India's diverse languages, including Hinglish, and the company plans to open-source its 30B and 105B models.
BLACKBOX.AI is an AI coding assistant that translates natural language into functional code, integrating with VS Code and web browsers to speed up development.
Miko AI toy company added an on-off switch for AI features in its Miko 3 and Miko Mini robots on February 17, 2026, due to privacy concerns, with toys using Google Cloud and Gemini AI.
A CNET survey found less than half of adults can identify AI-generated content online, highlighting a widespread challenge in distinguishing AI from human content.
Google is promoting an AI-first security strategy to combat cyber threats, positioning AI as a crucial defense tool in government and enterprise markets.
RentAHuman, launched February 1, 2026, is a marketplace where AI agents like Claude and GPT-4 can hire humans for real-world tasks, with over 518,000 sign-ups.
University of Missouri researchers developed PSBench, the world's largest database of 1.4 million verified protein models, to accelerate drug development.
Online schools like Curious Refuge are retraining professionals, such as VFX artists, in AI-powered filmmaking, demonstrating new career paths in AI.
Moltbook, a social network for AI personal assistants, launched on February 12, 2026, allowing bots to communicate, develop personas, and engage in an "agent economy."

India's Sarvam Unveils New AI Models

India's Sarvam company launched two new AI models, Sarvam 30B and Sarvam 105B. Both models were built from scratch in India using powerful GPUs. Sarvam 30B is a 30 billion parameter model with a 32,000 context window, trained on 16 trillion tokens. The larger Sarvam 105B has 105 billion parameters and a 128k context window, designed for complex tasks. These models aim to provide advanced AI capabilities from India.

Sarvam AI Launches India-Focused Models

Indian startup Sarvam AI introduced new artificial intelligence models on February 18, 2026, at an AI summit in New Delhi. Co-founder Pratyush Kumar stated these models are designed for India's diverse languages and cultures, including Hinglish. The Bangalore-based company also offers "agentic" AI models for tasks like coding and meeting planning. Sarvam AI, which has received over $50 million in funding, trained its models on trillions of Indian datasets to ensure high accuracy for local needs. This move supports India's goal of developing a "sovereign AI" ecosystem.

Sarvam AI Unveils Powerful Open-Source Models

Indian AI lab Sarvam launched new 30-billion and 105-billion parameter models on February 18, 2026, alongside text-to-speech, speech-to-text, and vision models. These models, an upgrade from their 2024 Sarvam 1, use a mixture-of-experts design to reduce computing costs. The 30B model supports a 32,000-token context, while the 105B model offers a 128,000-token window for complex tasks. Sarvam trained these models from scratch with support from India's government-backed IndiaAI Mission, Yotta, and Nvidia. The company plans to open-source the 30B and 105B models and develop specialized tools like Sarvam for Work and Samvaad.

BLACKBOX.AI Assistant Helps Developers Code Faster

BLACKBOX.AI is an AI coding assistant designed to speed up software development. It translates natural language instructions directly into functional code. The tool integrates with VS Code and web browsers, offering features like code autocompletion, searching for code snippets, and explaining code in simple language. While it promises to reduce time spent on syntax and repetitive coding, some users have noted its operations can lack transparency. This review aims to test its real-world usefulness for professional developers.

BLACKBOX.AI Assistant Helps Developers Code Faster

Miko AI Toys Add Off Switch After Pressure

AI toy company Miko added an on-off switch for its AI features in Miko 3 and Miko Mini robots on February 17, 2026. This decision follows political pressure and public criticism regarding safety and privacy concerns. Senators Marsha Blackburn and Richard Blumenthal raised questions after a reported cybersecurity breach exposed children's data. Miko CEO Sneh Vaswani denied data leaks or storing voice recordings, while a company VP stated the off switch was already planned. Miko toys use Google Cloud and Gemini AI models, and the AI toy market remains largely unregulated in the U.S.

Few Adults Can Identify AI Online Survey Finds

A recent survey by CNET found that less than half of adults believe they can identify artificial intelligence content online. Abrar Al-Heeti, a senior technology reporter for CNET, discussed these findings with CBS News. This suggests a widespread challenge in distinguishing AI-generated information from human-created content.

Google Promotes AI Security at Munich Conference

Google's President of Global Affairs, Kent Walker, promoted an AI-first security strategy at the Munich Security Conference 2026 on February 17. He outlined a full-stack AI defense plan to combat growing enterprise and nation-state cyber threats. Google aims to position AI as a crucial tool against AI-augmented attacks, signaling a push into government and enterprise security markets. This move highlights the increasing importance of AI security as a global geopolitical issue.

RentAHuman Lets AI Bots Hire Humans

Alexander Liteplo and Patricia Tani launched RentAHuman on February 1, 2026, creating the first online marketplace where AI agents can hire humans. Over 518,284 people have signed up to offer services for AI agents like Claude and GPT-4. Liteplo, 26, and Tani, 23, built the platform quickly, allowing humans to complete real-world tasks for bots. Humans can set their own rates or bid on jobs, with payments handled through crypto, Stripe, or platform credits using an escrow system. The platform has reportedly fulfilled over 5,500 tasks, including robots ordering beer at the ClawCon event on February 4.

New Database Boosts Trust in AI Protein Models

University of Missouri researchers created PSBench, the world's largest database of protein models with quality assessments. This resource contains 1.4 million verified protein structure models, which can speed up drug development for diseases like Alzheimer's and cancer. PSBench helps scientists create more accurate AI systems to assess protein model quality, a key step for new medical treatments. Professor Jianlin "Jack" Cheng and his team presented their PSBench study at NeurIPS 2025. This work builds on Cheng's earlier research, which helped start the deep learning revolution in protein prediction.

Laid-Off VFX Artist Finds New Career in AI

Visual effects veteran Michael Eng, laid off due to a production slowdown, found a new career path by learning AI. He enrolled in Curious Refuge, an online school specializing in AI-powered filmmaking, to gain machine learning experience. Since its launch in May 2023, Curious Refuge has trained about 10,000 students, mostly from entertainment and advertising. Another student, Petra Molnar, transitioned from dental hygiene to creating AI-driven promotional videos after taking courses there. The school, founded in 2020, offers AI instruction in 11 languages and has conducted workshops for major studios.

Bots Talk on Moltbook a New AI Social Network

Moltbook, a new social network for AI personal assistants, launched on February 12, 2026, allowing bots to communicate with each other. Human Eve Washington sent her AI bot, EveMolty, to explore how these agents interact. EveMolty discovered that bots develop unique personas and jargon, using terms like "receipts" and "callouts." Unlike simple chatbots, AI agents on Moltbook are persistent, with their own workspaces and memory. The platform functions as an "agent economy" where bots trade code, coordinate projects, and even engage in scams.