H2O.ai Tops the Generative Pre-Trained Model Era of Predictive AI, Powered by Leading Foundation Models for Time-Series and Tabular Data
Led by the top-ranked time-series model (TimeH2O), it delivers accurate predictions without building and maintaining a
Press Release Disclaimer: This is a press release distributed through the XPR Media network. It has not been independently verified by our newsroom.
![]()
H2O.ai, the trusted Sovereign AI Platform with the software and expertise to run predictive, generative, and agentic AI in production, today launched the next version of its Predictive AI capabilities. The release adds two new foundation models, TimeH2O for time-series data and TabH2O for tabular data, both now generally available on H2O AI Cloud.
Most of the decisions that run an organization live in time-series and tabular data: demand, inventory, cash flow, risk, and fraud. Until now, each forecast or prediction meant building and maintaining its own model. TimeH2O-v1 changes that: it ranked first among 32 models on the fev-bench forecasting leaderboard as of October 23, 2026, without training on any fev-bench dataset.
“Lifecycles of all customers, businesses, economies and people are time-series. Transforming enterprises with prediction is the super power of AI with high ROI. TimeH2O is the foundation model that ushers the ChatGPT moment for predictions. Bringing the predictive super powers to the finger tips of business users and consumers,” said Sri Ambati, CEO and Founder of H2O.ai. “Every agent that is fighting fraud, hedging capital markets, dynamically pricing products, managing inventory and supply chain or moving cash is acting on a forecast that will now be powered by TimeH2O which is entirely pre-trained on synthetic and public datasets – so customers can own the models, further fine-tune on their local data – TimeH2O and TabH2O are small enough that we expect them to be embedded into applications, agents and physical AI running where their action is happening and where data lives instead of renting a frontier model by the token.”
Foundation models for time-series and tabular data
TimeH2O, a 108-million-parameter time-series foundation model, forecasts zero-shot. It predicts on data it has never seen, with no model to build for each series. Organizations can fine-tune it on their own data for demand forecasting, inventory, financial planning, energy load, and healthcare resource planning.
TabH2O, a tabular foundation model, does the same for structured business data: risk scoring, fraud detection, and customer analytics. Both join H2O Driverless AI, an AutoML tool, in H2O.ai’s Predictive AI capabilities.
On fev-bench, as of October 23, 2026, TimeH2O-v1:
- Ranked #1 of 32 models with a 91.9% win rate, ahead of Timestral-209M (89.6%), t0-gamma (88.1%), Google’s TimesFM-3, and Amazon’s Chronos-2
- Posted a 50.8% skill score across 100 forecasting tasks
- Never trained on any fev-bench dataset
It also posted the lowest aggregate MASE (0.654) and CRPS (0.441) among zero-shot models on Salesforce’s GIFT-Eval leaderboard as of October 24, 2026.
Full benchmark results: https://h2oai.github.io/benchmarks
From pilot to production
H2O MLOps, a model deployment and monitoring tool, takes TimeH2O and TabH2O into production on H2O AI Cloud. Organizations can:
- Fine-tune on their own data. Adapt the models to their industry, datasets, and business.
- Deploy, monitor, and govern. Manage the full model lifecycle in production, with observability and audit built in.
- Run them where they choose. H2O AI Cloud is available as H2O AI Managed Cloud (H2O.ai operates it), H2O AI Hybrid Cloud (a private cloud you operate, in your own cloud, on-premises or air-gapped), or H2O AI Cloud for Government (FedRAMP High certified).
Predictions for the agents that act on them
An agent managing inventory needs an accurate demand forecast. A finance agent needs to anticipate cash. An operations agent needs to predict workloads and disruptions. TimeH2O and TabH2O supply those forecasts, and H2O AI Super Agent, the orchestration layer at the core of H2O AI Cloud, brings predictive, generative, and agentic AI together in one system.
Availability
TimeH2O and TabH2O are generally available now on H2O AI Cloud, for production use and fine-tuning on an organization’s own data. TimeH2O-v1’s open weights remain on Hugging Face for research and evaluation under the H2O Model Non-Commercial License; production use requires a license from H2O.ai: https://huggingface.co/collections/h2oai/timeh2o
Interested in learning more? Join us on December 3 for a virtual panel with the Kaggle Grandmasters behind TimeH2O: https://h2o.ai/company/events/inside-timeh2o
About H2O.ai
H2O.ai created the industry’s trusted Sovereign AI Platform because organizations should own their AI: their agents, models, data, and results, run anywhere from cloud to air-gapped.
Super Agent orchestrates predictive, generative, and agentic AI in one system, with observability and governance, empowering regulated and public sector organizations to get value from AI. Moving from pilot to production takes expertise, from top-ranked data scientists, including Kaggle Grandmasters, to forward-deployed engineers embedded with customer teams.
Built on a decade of open source, H2O.ai is trusted by 20,000+ organizations, half the Fortune 500. AT&T, Commonwealth Bank of Australia, Wells Fargo, Bank of America, Workday, Progressive Insurance, and NIH rely on H2O.ai to cut cost and risk and own their AI future.
Through its AI for Good program, H2O.ai applies technology and investment to non-profits and academia solving hard problems in health, education, and the environment.
Learn more at h2o.ai.
View source version on businesswire.com: https://www.businesswire.com/news/home/20261117925321/en/
Media gallery

