H2O.ai launches TimeH2O-v1 forecasting model, tops fev-bench

H2O.ai has released TimeH2O-v1, a time-series foundation model for forecasting, and says it ranks first on the fev-bench leaderboard by win rate.

According to H2O.ai, TimeH2O-v1 placed first among 32 models in the benchmark's default view. It posted a 91.9% win rate and a skill score of 50.8 across 100 forecasting tasks, ahead of Timestral-209M, t0-gamma, TimesFM-3 and Chronos-2.

The release puts H2O.ai into direct competition with larger technology groups and specialist model developers in a part of the artificial intelligence market focused on operational prediction rather than text generation. Time-series forecasting is used in demand planning, inventory management, cash forecasting, energy load management and patient volume planning.

Unlike forecasting systems trained separately for each dataset, TimeH2O-v1 works in zero-shot mode, H2O.ai said. This means it reads the history of a time series and generates a forecast without training on that specific dataset at the point of use.

H2O.ai said TimeH2O-v1 was not trained on any fev-bench dataset, and that no training or fine-tuning takes place during forecasting, leaving the model's weights unchanged. The benchmark measures probabilistic accuracy on real-world time series across different domains and frequencies against a seasonal-naive baseline.

Benchmark results

On the leaderboard view cited by H2O.ai, TimeH2O-v1 beat rival models in 91.9% of head-to-head comparisons. Timestral-209M followed at 89.6%, while t0-gamma recorded 88.1%.

H2O.ai said the model's skill score of 50.8 means forecast error was 50.8% lower than the seasonal-naive baseline when aggregated across all 100 tasks.

TimeH2O-v1 has 108 million parameters, according to H2O.ai, and can forecast both univariate series and series that include covariates such as price, promotions or weather.

H2O.ai said the model performs across datasets from finance, healthcare, supply chain and energy. It added that organisations can fine-tune it on their own infrastructure using their own data to improve results for specific domains.

Business focus

The launch reflects a growing effort among AI developers to build models for narrower business tasks, where prediction accuracy can be measured more directly than in general-purpose chatbots. Forecasting has long been a core use case for machine learning in large organisations, but many companies still maintain separate models for separate operational problems and retrain them as conditions change.

H2O.ai is positioning TimeH2O-v1 as a general model that can serve as a starting point before adaptation to a customer's data. The approach mirrors a wider industry shift towards foundation models tailored to specific data types, including time series, code, biology and industrial processes.

"Building your business with accurate forecasting and prediction is the superpower of AI. TimeH2O gives organizations a generalized, pre-trained model for time series that can then be fine-tuned into a domain-specific foundation model for each customer's use case," said Sri Ambati, Chief Executive Officer and Founder of H2O.ai.

TimeH2O-v1 is available under a non-commercial licence for internal benchmarking and testing. Organisations with commercial forecasting requirements would need separate arrangements, H2O.ai said.

H2O.ai said it is used by more than 20,000 organisations, including half of the Fortune 500. It cited customers including AT&T, Commonwealth Bank of Australia, Wells Fargo, Bank of America, Workday, Progressive Insurance and the NIH.