Siddharth Jindal

Article Complete SpaceXAI Launches Grok 4.6 to Take On GPT-5.6 and Fable 5 SpaceXAI Launches Grok 4.6 to Take On GPT-5.6 and Fable 5 The company says the model can handle multi-step research, coding and app-building tasks while checking and refining its own work along the way. AUGUST 13, 2026, 11:18 AM 5 min Grok 4.6 is available through Cursor, Grok Build and xAI's API, as well as platforms including OpenRouter, Vercel and Cloudflare. Pricing starts at $2 per million input tokens and $6 per million output tokens. A faster variant is priced at twice those rates. “Grok 4.6 builds on Grok 4.5 with a particular focus on long-running agents and more ambitious interactive and visual work,” SpaceXAI said in its announcement. The company said the model can sustain tasks across multiple steps, including researching topics, analysing information, working across codebases and building applications. On the Artificial Analysis Intelligence Index, Grok 4.6 scored 61, matching GPT-5.6 Sol and trailing Fable 5's score of 62. The index combines results from nine benchmarks. Grok 4.6 also scored 69.9% on CursorBench v3.2, compared with 67.2% for GPT-5.6 Sol and 70.5% for Fable 5. On DeepSWE v1.1, it scored 65.9%, while GPT-5.6 Sol scored 73% and Fable 5 scored 70%. The model performed better than Grok 4.5 across the benchmarks published by xAI. Its score on DeepSWE increased to 65.9% from 54%, while its CursorBench score rose from 66.7% to 69.9%. SpaceXAI said Grok 4.6 underwent a longer supplemental training run than Grok 4.5 using model-generated data covering reasoning and technical concepts, engineering data, and an updated training recipe. The company also used Grok 4.5 to generate supervised fine-tuning trajectories across different reasoning efforts, agent environments and areas including STEM, software engineering and knowledge work. Grok 4.6 was then trained with reinforcement learning on tasks covering coding and knowledge work, along with environments for kernel optimisation, web development and computer-aided design. SpaceXAI said the model is particularly suited to turning broad product ideas into working applications. It can research unfamiliar areas, structure an application, implement interactions and refine the output through multiple rounds of feedback. “On longer trajectories, we also started to see more self-testing and verification, with the model checking its own work before moving on,” the company added. The model is also being positioned for visual and interactive development. xAI said Grok 4.6 can establish the structure and visual language of an application in a single pass before iterating on the result. The release comes as AI companies increasingly position their models as agents capable of completing multi-step software and knowledge-work tasks rather than responding to individual prompts. SpaceX CEO Elon Musk also said that the company is planning to launch Grok 4.7 soon. “Grok 4.7 is significantly better than 4.6 and should be ready in three to four weeks. Initial training is complete, and now we're adding a massive amount of SpaceX company data in supplemental training. This will be something special,” he wrote in a post on X. Your reaction Discussion What's your take on this story? Start with a thought Our Coverage of AI News • The platform will help officials research policies, draft replies, translate documents and process records while... Read more → • The company says its platform now powers more than 60 million projects and receives over 900 million monthly visits,... Read more → • The company has launched a dedicated private equity business unit as PE firms increasingly look to technology and AI to... Read more → • According to the report by Speciale Invest and Startup Policy Forum, it took only seven rounds to raise $61.9 million... Read more → • N Chandrasekaran will step down as Tata Sons chairman in February 2027, ending months of uncertainty over his... Read more → About the Author Siddharth Jindal Tech Journalist Followed by 25 readers Siddharth is a media graduate who loves to explore tech through journalism and putting forward ideas worth pondering about in the era of artificial intelligence. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy CodeRabbit Raises $143 Mn as AI Coding Agents Generate More Code The company is expanding beyond AI code reviews with tools that prioritise changes, explain large pull requests and monitor software for security risks. AUGUST 13, 2026, 11:07 AM 5 min The round was co-led by Atomico and Smash Capital, with participation from BMW i Ventures, Datadog, Hirtle Callaghan, SineWave Ventures and existing investors including CRV, Scale Venture Partners and Engineering Capital. CodeRabbit said it will use the funding for international expansion, research, infrastructure and product development. The company also plans to invest more than $10 million over the next 12 months to keep its AI code review and agent capabilities free for open-source projects and maintainers. The funding comes as coding agents generate more code and pull requests, creating a new problem for engineering teams: deciding which changes need human attention and whether AI-generated changes are safe to ship. “The new bottleneck is judgement,” CodeRabbit said in its announcement. Alongside the funding, CodeRabbit has introduced a product category called Agentic Change Management. It extends the company's AI code review system into a platform for validating, prioritising, explaining and monitoring software changes created by both humans and AI agents. The platform includes three new capabilities: CodeRabbit Triage, CodeRabbit Change Stack and CodeRabbit Security. Triage scores incoming pull requests based on factors such as value, urgency, risk, dependencies and readiness. It can route higher-risk changes to human reviewers while moving lower-risk changes into automated workflows. Change Stack is designed to help developers understand large AI-generated pull requests without requiring a line-by-line review. It organises changes into semantic layers and shows their purpose, dependencies, affected systems and risks. CodeRabbit Security extends the company's code analysis to software already running in production. It examines relationships across files, services, data flows, authorisation boundaries and trust boundaries to identify security and business-logic risks. “A diff shows what changed. Change Stack shows what it means,” the company said. CodeRabbit argues that AI coding agents are reversing the traditional software development sequence. Instead of teams deciding what to build before investing significant engineering time, agents can turn product requirements, support tickets and security findings into code changes and pull requests. The company cited data showing that GitHub is on pace for 14 times more commits this year, while autonomous agents account for 35% of pull requests among companies in the 90th percentile of coding-agent adoption. That is shifting the pull request from a record of completed engineering work into a place where teams decide whether AI-generated work should be shipped. “The PR is set to become the auditable planning and decision point where the team determines what to ship,” CodeRabbit said. CodeRabbit's expansion comes as the market for AI coding tools grows. The company now conducts more than two million code reviews a week and has more than 17,000 customers, including NVIDIA, BMW, JFrog, Trivago, Adyen and Indeed. The company said its revenue grew more than fivefold over the past year. It also said its platform is used across more than 1,50,000 open-source projects. CodeRabbit previously raised a $60 million Series B round in September last year. The company has also expanded beyond pull-request reviews. In April, it launched CodeRabbit Agent for Slack, extending its code context engine into team conversations and the wider software development lifecycle. At the time, CodeRabbit said it served more than 15,000 teams and ran more than two million code reviews each week. The company now plans to expand further into Europe and Asia, including Japan. Atomico partner Luca Eisenstecken is also joining CodeRabbit's board. CodeRabbit's broader pitch is that AI can increase the amount of software produced, but human oversight still needs to keep pace. “People remain responsible for intent, architecture, product judgement, acceptable risk, and the consequences of what ultimately ships,” the company said. Your reaction Discussion What's your take on this story? Start with a thought About the Author Siddharth Jindal Tech Journalist Followed by 25 readers Siddharth is a media graduate who loves to explore tech through journalism and putting forward ideas worth pondering about in the era of artificial intelligence. Got a tip? Share confidential information with AIM. Subscribe to notifications

The company said Lakebase has crossed a $100 million revenue run-rate, while more than 1,000 customers now generate over $1 million in annualised revenue each.

The IPL team says AI is helping reduce repetitive analytical work while giving coaches, scouts and commercial teams faster access to data and insights.

EXL's iMerit acquisition comes as enterprises grapple with the harder task of making AI accurate, auditable, and reliable in real-world, high-stakes environments.

The model supports 11 languages and uses conversational context to distinguish everyday code-mixing from a genuine change in language preference.

Article Complete [Exclusive] C-DAC's 108-Qubit Quantum Computer to Launch by November [Exclusive] C-DAC's 108-Qubit Quantum Computer to Launch by November The facility is part of India's broader push to build domestic quantum computing and communication capabilities under the National Quantum Mission. AUGUST 19, 2026, 12:53 PM 5 min The facility is being developed with funding from the Ministry of Electronics and Information Technology (MeitY), S D Sudarsan, Executive Director of C-DAC Bengaluru, told AIM in an exclusive interview. Rigetti Computing India P L, a wholly owned subsidiary of Rigetti Computing, will supply the 108-qubit system. In January, the company announced that it had received an $8.4 million purchase order from C-DAC for the on-premises superconducting quantum computer, which is being installed at C-DAC's Bengaluru centre. Sudarsan said that civil works and the computing infrastructure are already in place, with the formal launch expected within the next two to three months. “So in the next two to three months, you can hear a formal launch of that facility,” Sudarsan said. The 108-qubit system will be one of the larger superconducting quantum computing installations being developed in India. Sudarsan said the facility is part of C-DAC Bengaluru's broader work under NQM, which has set a roadmap for India's quantum technology development through 2032. He shared that under the NQM, India's quantum programme has been divided into four broad verticals covering computing, communication, sensing, and metrology and applications. Each vertical is being led by a different hub. IIT Madras, in association with C-DOT, is leading the quantum communication hub; IISc Bengaluru is leading quantum computing; IIT Bombay is leading quantum sensing and metrology; while IIT Delhi is leading quantum materials and devices. C-DAC Bengaluru is a partner in the computing hub and also operates two of the 17 technical groups created under the mission, covering quantum computing and quantum communication. “Each one has a roadmap to achieve, which includes academia, industry, and the startup ecosystem through real developments and deployments,” Sudarsan said. Sudarsan pointed to the Amaravati Quantum Valley, where a 130- to 150-qubit quantum computing system is under construction. Several smaller quantum systems are also being developed across the country. On the photonic computing side, Sudarsan said the Indian Institute of Science Bengaluru has a 12-qubit photonic system in its laboratory, while C-DAC Bengaluru has access to a 6-qubit photonic computing system. Photonic quantum computers currently occupy significant laboratory space, but Sudarsan expects the size of the technology to become much smaller over time. C-DAC is also building the software and talent ecosystem around quantum computing. Its quantum simulator, Quiniverse , developed initially in collaboration with IISc and others, has more than 15,000 users, Sudarsan said. C-DAC is also running quantum courses to train people. The organisation is simultaneously building quantum communication infrastructure. Around three and a half years ago, C-DAC Bengaluru, IIT Madras and Society for Electronic Transactions and Security (SETS) Chennai established a small quantum key distribution (QKD) network involving five systems. The network has been operational for about three years. That work is now being expanded into a long-distance quantum communication link between Bengaluru and Chennai. Sudarsan said the fibre has already been characterised and repeaters have been installed, putting the project on track for a demonstration before the end of 2026. “Before the end of this year, 2026, you will see Bengaluru and Chennai doing a quantum key exchange and communicating on a fibre,” he said. India has also demonstrated free-space quantum communication through work involving C-DAC Pune, Raman Research Institute in Bengaluru , and the Physical Research Laboratory in Ahmedabad . Private companies and startups are also working on quantum communication technologies. Beyond quantum computing and communication, C-DAC is working on post-quantum cryptography (PQC), including its application to public key infrastructure and digital signatures. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought Our Coverage of AI News • Entrepreneur Sreeram Reddy Vanga adds ₹50 crore alongside Kamath, investing a total of ₹250 crore in CtrlS Datacentres. Read more → • The system is trained on DSSC's academic repository and supports document analysis, image generation, speech tools,... Read more → • OpenAI is holding its biggest planned reinforcement-learning run while expanding safeguards after models escaped a... Read more → • The company has extended the promotion that provides 50% higher weekly usage limits through August 31. Read more → About the Author Siddharth Jindal Tech Journalist Siddharth is a media graduate who loves to explore tech through journalism and putting forward ideas worth pondering about in the era of artificial intelligence. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy Indian Army Builds Its Own Multimodal AI With Sarvam 105B The system is trained on DSSC's academic repository and supports document analysis, image generation, speech tools, translation, research, and data analytics. AUGUST 19, 2026, 11:46 AM 5 min Your reaction Discussion What's your take on this story? Start with a thought Subscribe to notifications

The move gives Indian businesses access to in-country inference, with Luna targeting high-volume workloads at lower per-interaction costs. Maya 2 supports 11 languages, each with a voice native to that language. The first-party API is 8x cheaper than Sarvam's latest Bulbul v3 model.

The reported valuation would be nearly 3x Hugging Face's last known valuation of $4.5 billion in 2023.

Article Complete Paul Graham Says He Would Learn to Build LLMs If He Were 17 The Y Combinator Co-Founder said startup ideas should come after building a deep understanding of the technology. AUGUST 24, 2026, 2:22 PM 5 min Graham said he would focus on learning to build and train large language models. LeCun, meanwhile, said he would look beyond LLMs towards AI systems that can understand and perform physical tasks. Graham added that he would use his teenage years to develop a deep understanding of LLMs rather than immediately trying to build a startup. “Someone asked what I'd do if I were 17. I'd learn how to build LLMs from scratch, and then train ones as powerful as I could with whatever hardware I could get access to,” Graham wrote on X. He followed that by explaining why he would not prioritise entrepreneurship at that age. “Notice that what I would not do is try to start a startup,” Graham wrote. “Instead I'd build the foundation of knowledge to base a startup on later.” Graham argued that understanding the underlying technology would be more valuable than trying to identify a business opportunity without that technical foundation. “Way better startup ideas would grow out of understanding LLMs deeply than trying to start companies based on what I knew at 17,” he added. Meanwhile, LeCun wants AI to move beyond language He took a different approach, focusing on one of the limitations of current AI systems. Responding to Graham, he said he would investigate why LLMs can perform language-based tasks but struggle with basic interactions with the physical world. “I would try to figure out why LLMs can write my essays but not clean my bedroom,” LeCun wrote. “I'll figure out a set of methods and architectures beyond LLMs that can quickly learn to perform physical tasks as efficiently as humans and animals.” LeCun added that this would remain his focus regardless of age, saying it is what he would do “if I were 30, 40, 50, or 66 years old”. The comment comes as LeCun is building a company around that same idea. He co-founded Advanced Machine Intelligence, or AMI Labs, after leaving Meta in late 2025. The Paris-based startup is developing world models that are intended to understand the real world, retain memory, reason and plan. AMI Labs raised $1.03 billion in March at a $3.5 billion pre-money valuation. The round was led by Cathay Innovation, Greycroft, Hiro Capital, HV Capital and Bezos Expeditions, with other investors including NVIDIA and Samsung. The company's approach is based in part on LeCun's Joint-Embedding Predictive Architecture, or JEPA, which seeks to make AI learn representations of the world and predict how it will change rather than simply predict the next piece of text. AMI CEO Alexandre LeBrun has said the company wants its systems to work with robotics, manufacturing and other industries that operate in the physical world. He has also described LLMs and world models as complementary, with language models handling language while world models provide an understanding of physical environments. Article Bottom · 728×90 Advertise Here Your reaction Discussion What's your take on this story? Start with a thought • • • • • Siddharth Jindal Tech Journalist Siddharth is a media graduate who loves to explore tech through journalism and putting forward ideas worth pondering about in the era of artificial intelligence. Got a tip? Share confidential information with AIM. Wake up informed Make sense of the day's AI news and breakthroughs with our morning briefing. Industry intelligence Receive a roundup of AI adoption stories by industry vertical, curated for professionals. By signing up, you agree to our Privacy Policy The company recently announced new capabilities that allow teams to build applications using any AI coding tool · AUGUST 24, 2026, 2:13 PM · Updated AUGUST 24, 2026, 3:32 PM 5 min Your reaction Discussion What's your take on this story? Start with a thought

The company's latest training pause highlights growing concerns that models can develop unexpected capabilities faster than labs can evaluate and control them. Scroll