H2O.ai Breaks New World Record for Most Accurate Agentic AI for Generalized Assistants
MOUNTAIN VIEW, Calif.--(BUSINESS WIRE)-- H2O.ai, the #1 Agentic AI company, announced its h2oGPTe Agent has once again achieved the top ranking on the prestigious General AI Assistant (GAIA) benchmark, with a record-setting 79.7% accuracy, fast approaching human-level performance, measured at 92%, and far surpassing general-purpose models from Google and Microsoft, which scored below 50%.
GAIA is a rigorous evaluation framework that measures how effectively AI agents perform over 300 real-world tasks, spanning research, data analysis, document handling, and advanced reasoning. It serves as a key indicator of enterprise readiness, assessing whether AI systems can handle the kinds of high-effort, skilled tasks traditionally done by humans. For business leaders, this milestone means h2oGPTe is capable of handling the kind of nuanced decisions, regulations, and workflows that drive value in banking, telecom, healthcare, and government. H2O.ai first achieved the top GAIA ranking in December 2024 and continues to lead in production-grade, sovereign AI.
The improved performance of H2O.ai's agent technology comes from key enhancements, including advanced browser navigation for precise information extraction, unified search across multiple sources like Google and Wikipedia, and the integration of Google's Gemini 2.5 Pro and Claude 3.7 Sonnet. Additionally, the platform now features GitHub integration for navigating codebases and real-time source attribution, ensuring transparency during research.
'GAIA is fast becoming the barometer of enterprise intelligence, and at 79.7%, our agents aren't just accurate, they're adaptable,' said H2O.ai CEO and Founder Sri Ambati. 'Gemini sharpened our vision and multimodal skills, Claude boosted our reasoning and code understanding, and now we're building toward an auto-agentic future, a framework where planning agents coordinate a series of task-specific power tools. DeepResearch already gave hedge funds an edge in volatile markets, and in today's shifting geopolitical landscape, scenario planning is not a luxury, it's a necessity. Delivering all of this, on-prem, inside sovereign AI environments for governments and public institutions, that's a game changer.'
H2O.ai's agents are deployed in some of the world's most highly regulated environments to support mission-critical, task-specific operations. Global banks use them to streamline regulatory reporting and detect fraud, telecom providers to optimize call centers, and public agencies to manage complex document workflows. H2O.ai offers a growing portfolio of vertical agents — prebuilt for industries like banking, telecom, and government — and a flexible agent builder framework for creating custom agents on private data and internal systems. Built on a multi-agent architecture, planning agents can coordinate specialized sub-agents across departments, delivering structure, speed, and scale. With human-in-the-loop review, continuous learning, and auditability built in, H2O agents meet strict compliance needs while accelerating decision-making and ROI.
As enterprises move from AI pilots to production, H2O.ai continues to lead the way with modular, hardware-agnostic solutions that run securely on private clouds, on-premise infrastructure, or air-gapped environments.
With GAIA as a clear signal of applied intelligence, H2O.ai stands apart in an increasingly crowded field, proving what's possible when agentic AI is purpose-built for the enterprise.
For more information about H2O.ai's capabilities, visit h2o.ai.
About H2O.ai
Founded in 2012, H2O.ai is on a mission to democratize AI. As the world's leading agentic AI company, H2O.ai converges Generative and Predictive AI to help enterprises and public sector agencies develop purpose-built GenAI applications on their private data. With a focus on Sovereign AI—secure, compliant, and infrastructure-flexible deployments—H2O.ai delivers solutions that align with the highest standards of data privacy and control.
Its open-source technology is trusted by over 20,000 organizations worldwide, including more than half of the Fortune 500. H2O.ai powers AI transformation for companies like AT&T, Commonwealth Bank of Australia, Singtel, Chipotle, Workday, Progressive Insurance, and NIH.
H2O.ai partners include Dell Technologies, Deloitte, Ernst & Young (EY), NVIDIA, Snowflake, AWS, Google Cloud Platform (GCP) and VAST. H2O.ai's AI for Good program supports nonprofit groups, foundations, and communities in advancing education, healthcare, and environmental conservation. With a vibrant community of 2 million data scientists worldwide, H2O.ai aims to co-create valuable AI applications for all users.
H2O.ai has raised $256 million from investors, including Commonwealth Bank, NVIDIA, Goldman Sachs, Wells Fargo, Capital One, Nexus Ventures and New York Life.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
22 minutes ago
Federal judge sides with Meta in AI copyright case
SAN FRANCISCO -- Federal judge sides with Meta in AI copyright case but leaves door open for similar lawsuits.
Yahoo
24 minutes ago
- Yahoo
Overbought Micron Stock Could Be Volatile After Q3 Earnings
Micron Technology (MU) popped higher in after-hours trading following its exceptional fiscal third-quarter 2025 results - but the stock has quickly pared the bulk of its late gains, and is up just 0.5%. Heading into tonight's report, MU stock had surged more than 36% over the past month, and the 14-day Relative Strength Index (RSI) of 83 was deep in overbought territory. That means the shares could be vulnerable to a short-term technical reversal, despite the strong earnings. Looking at the results, Micron reported record revenue of $9.3 billion, marking a substantial 37% year-over-year increase. This impressive growth was primarily driven by surging demand for artificial intelligence (AI)-related memory products, and was particularly evident in the company's data center segment, where revenue more than doubled compared to the previous year. Dear Micron Stock Fans, Mark Your Calendars for June 25 Is United Health Stock a Buy, Hold or Sell for July 2025? This New ETF Promises to Help You Invest Like Warren Buffett and Yields 15% Tired of missing midday reversals? The FREE Barchart Brief newsletter keeps you in the know. Sign up now! The company's DRAM segment, accounting for 76% of total revenue, reached an all-time high of $7.1 billion, representing a 51% year-over-year increase. High-bandwidth memory (HBM) revenue grew nearly 50% sequentially, while NAND flash revenue showed more modest growth, contributing $2.2 billion to the total. Micron's adjusted earnings per share (EPS) of $1.91 significantly exceeded analysts' expectations of $1.60, while gross margins expanded to 39%, demonstrating improved operational efficiency. The Compute and Networking Business Unit showed remarkable strength with $5.1 billion in revenue, up 97% year-over-year, primarily driven by robust data center demand for AI applications. Looking ahead, Micron provided strong guidance for its fiscal fourth quarter, projecting revenue between $10.4-11.0 billion and adjusted earnings per share of $2.35-2.65, well above consensus estimates. The company's strategic positioning in the AI memory market appears particularly strong, with HBM now shipping to four high-volume customers and development underway for next-generation HBM4 chips targeted for 2026 production. Management expressed confidence in the company's outlook, citing tight DRAM inventories and improving NAND inventory levels as positive indicators for continued market momentum. However, some analysts have noted potential risks from tariff-related inventory stockpiling that could impact future quarters. Despite these concerns, Micron's strong free cash flow generation of $1.95 billion demonstrates its ability to fund continued investments in advanced memory technologies while maintaining operational efficiency. This article was generated with the support of AI and reviewed by an editor. On the date of publication, the editor did not have (either directly or indirectly) positions in any of the securities mentioned in this article. All information and data in this article is solely for informational purposes. This article was originally published on Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
Yahoo
31 minutes ago
- Yahoo
Meta fends off authors' US copyright lawsuit over AI
By Blake Brittain (Reuters) -A federal judge in San Francisco ruled on Wednesday for Meta Platforms against a group of authors who had argued that its use of their books without permission to train its artificial intelligence system infringed their copyrights. U.S. District Judge Vince Chhabria said in his decision the authors had not presented enough evidence that Meta's AI would dilute the market for their work to show that the company's conduct was illegal under U.S. copyright law. Chhabria also said, however, that using copyrighted work without permission to train AI would be unlawful in "many circumstances," splitting with another San Francisco judge who found on Monday in a separate lawsuit that Anthropic's AI training made "fair use" of copyrighted materials. "This ruling does not stand for the proposition that Meta's use of copyrighted materials to train its language models is lawful," Chhabria said. "It stands only for the proposition that these plaintiffs made the wrong arguments and failed to develop a record in support of the right one." Spokespeople for Meta and attorneys for the authors did not immediately respond to requests for comment. The authors sued Meta in 2023, arguing the company misused pirated versions of their books to train its AI system Llama without permission or compensation. The lawsuit is one of several copyright cases brought by writers, news outlets and other copyright owners against companies including OpenAI, Microsoft and Anthropic over their AI training. The legal doctrine of fair use allows the use of copyrighted works without the copyright owner's permission in some circumstances. It is a key defense for the tech companies. Chhabria's decision is the second in the U.S. to address fair use in the context of generative AI, following U.S. District Judge William Alsup's ruling on the same issue in the Anthropic case. AI companies argue their systems make fair use of copyrighted material by studying it to learn to create new, transformative content, and that being forced to pay copyright holders for their work could hamstring the burgeoning AI industry. Copyright owners say AI companies unlawfully copy their work to generate competing content that threatens their livelihoods. Chhabria expressed sympathy for that argument during a hearing in May, which he reiterated on Wednesday. The judge said generative AI had the potential to flood the market with endless images, songs, articles and books using a tiny fraction of the time and creativity that would otherwise be required to create them. "So by training generative AI models with copyrighted works, companies are creating something that often will dramatically undermine the market for those works, and thus dramatically undermine the incentive for human beings to create things the old-fashioned way," Chhabria said.
