Introducing TimesFM 3.0 with multivariate time-series forecasting in BigQuery

Google announced the preview of TimesFM 3.0 in BigQuery, a 330-million parameter foundation model from Google Research that adds native multivariate time-series forecasting to BigQuery SQL via AI.FORECAST. TimesFM 3.0 ranks #1 on all three major public time-series leaderboards — GIFT-Eval, FEV-Bench, and TIME — and serves as a drop-in upgrade for univariate forecasting and anomaly detection by specifying model => 'TimesFM 3.0'. The model supports multiple target columns and both past-only and past-and-future covariate columns, capabilities unavailable in TimesFM 2.0 and 2.5.

Since introducing zero-shot time-series forecasting in BigQuery with AI.FORECAST https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-forecast — initially powered by TimesFM 2.0 and later upgraded to TimesFM 2.5 — you have been able to generate forecasts directly in SQL without training or managing custom models. However, up until TimesFM 2.5, these foundation models were limited to univariate forecasting, predicting future values from the history of a single time series alone. In practice, most real-world forecasting problems are multivariate, where related targets and external drivers like weather, promotions, and holidays jointly shape future outcomes. Today, we are thrilled to announce the preview of TimesFM 3.0 in BigQuery. Developed by Google Research, TimesFM 3.0 https://research.google/blog/timesfm-3-a-zero-shot-foundation-model-for-multivariate-forecasting is a 330-million parameter state-of-the-art foundation model designed to unlock native multivariate time-series forecasting directly inside BigQuery SQL. TimesFM 3.0 ranks 1 across all three major public time-series leaderboards — GIFT-Eval, FEV-Bench, and TIME — achieving top zero-shot accuracy on both covariate and multivariate forecasting tasks compared to existing foundation models. Alongside its new multivariate capabilities, TimesFM 3.0 also serves as a drop-in upgrade for univariate forecasting via AI.FORECAST and anomaly detection via AI.DETECT ANOMALIES https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-detect-anomalies . Existing workflows can immediately benefit from higher benchmark accuracy simply by specifying model = 'TimesFM 3.0'. With TimesFM 3.0, BigQuery brings enterprise-grade multivariate modeling into familiar SQL workflows via AI.FORECAST: | Capability | TimesFM 2.0 / 2.5 | TimesFM 3.0 | | Forecasting scope | Univariate only | Univariate and multivariate | | Target columns | Single target | Single or multiple targets | | Covariate columns with past-only values | Not supported | Supported | | Covariate columns with past and future values | Not supported | Supported | To see how these capabilities work together, let’s forecast daily commuter subscriber trips and leisure casual trips bike rides over a seven-day horizon using the NYC Citi Bike public dataset bigquery-public-data.new york.citibike trips https://console.cloud.google.com/bigquery?ws= 1m5 1m4 4m3 1sbigquery-public-data 2snew york 3scitibike trips . BigQuery uses a single-table input format for multivariate forecasting that combines two sets of rows: Historical observation rows : timestamps between start time 2016-01-01 and the forecast cutoff date 2016-06-01 , fully populated with target metrics, past covariates, and known future covariates Future horizon rows : timestamps across the seven-day forecast horizon 2016-06-01 to 2016-06-07 populated with known future covariates, while target columns and past-only covariates are set to NULL This query highlights the three multivariate dimensions supported in TimesFM 3.0: Forecasting correlated targets target cols Instead of running separate models for each rider segment, passing TARGET COLS = 'subscriber trips', 'casual trips' forecasts both metrics jointly in a single call. TimesFM 3.0 captures the cross-series dynamics between commuter and casual riders — such as shared weather sensitivity and diverging weekend ridership patterns. Using past covariates past covariate cols Some signals can only be observed historically and are not known in advance, such as average ride duration avg duration min , historical foot traffic, or unplanned service disruptions. Passing PAST COVARIATE COLS = 'avg duration min' with future horizon rows set to NULL gives the model historical context to distinguish temporary shocks from underlying trends without requiring future values for that feature. Using future covariates future covariate cols Other external drivers are known ahead of time across both the historical window and the forecast horizon — such as weekends is weekend , holidays, or scheduled promotions. Passing FUTURE COVARIATE COLS = 'is weekend' populated across all rows through 2016-06-07 allows TimesFM 3.0 to condition its seven-day predictions directly on upcoming calendar events. The function returns structured predictions for each target column specified in target cols. For each column in target cols, AI.FORECAST returns a struct containing the point forecast value and the lower and upper prediction interval bounds: | date | subscriber trips.value | subscriber trips.prediction interval lower bound | subscriber trips.prediction interval upper bound | casual trips.value | casual trips.prediction interval lower bound | casual trips.prediction interval upper bound | time series type | confidence level | ai forecast status | | 2016-06-01 00:00:00 UTC | 44699.64 | 32590.76 | 55723.44 | 4581.32 | 1727.91 | 9189.31 | forecast | 0.95 | | | 2016-06-02 00:00:00 UTC | 44611.75 | 30413.05 | 57209.19 | 4647.17 | 1466.23 | 10273.55 | forecast | 0.95 | | | 2016-06-03 00:00:00 UTC | 41379.15 | 26218.65 | 54940.09 | 5341.36 | 1568.23 | 11805.86 | forecast | 0.95 | | | 2016-06-04 00:00:00 UTC | 30164.34 | 14266.43 | 45075.92 | 9097.74 | 2558.27 | 18452.52 | forecast | 0.95 | | | 2016-06-05 00:00:00 UTC | 26727.27 | 10101.44 | 42239.33 | 8398.50 | 1692.15 | 17768.99 | forecast | 0.95 | | | 2016-06-06 00:00:00 UTC | 36713.84 | 19493.00 | 52679.59 | 4723.08 | 400.08 | 11710.73 | forecast | 0.95 | | | 2016-06-07 00:00:00 UTC | 39749.97 | 22157.42 | 56119.86 | 4180.97 | 95.70 | 10845.74 | forecast | 0.95 | | You can also run multivariate forecasts using natural language in BigQuery Conversational Analytics. On the same bigquery-public-data.new york.citibike trips https://console.cloud.google.com/bigquery?ws= 1m5 1m4 4m3 1sbigquery-public-data 2snew york 3scitibike trips dataset, you can simply ask "Forecast both daily subscriber and casual trips for June 1–7, 2016 using two month historical data, with average trip duration as a past covariate and weekend indicator as a future covariate." Conversational Analytics automatically generates and executes the underlying AI.FORECAST query with TimesFM 3.0, returning a summary of the results, interactive forecast charts with prediction intervals for each target, and key forecast insights. With native multivariate support in TimesFM 3.0, BigQuery brings state-of-the-art foundation model forecasting to complex, interconnected enterprise datasets without requiring data movement, custom feature engineering pipelines, or dedicated model hosting infrastructure. To start forecasting with TimesFM 3.0 in BigQuery: Read the Google Research blog post https://research.google/blog/timesfm-3-a-zero-shot-foundation-model-for-multivariate-forecasting to learn more about the TimesFM 3.0 model architecture and capabilities. Explore the BigQuery TimesFM https://docs.cloud.google.com/bigquery/docs/timesfm-model documentation. Follow the step-by-step multivariate forecasting tutorial https://docs.cloud.google.com/bigquery/docs/timesfm-multivariate-multi-time-series-forecasting-tutorial . Try AI.FORECAST https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-forecast with model = 'TimesFM 3.0' directly in the BigQuery Studio SQL workspace. Try AI.DETECT ANOMALIES https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-detect-anomalies with model = 'TimesFM 3.0' for univariate time series anomaly detection in BigQuery. Try AI.EVALUATE https://docs.cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-ai-evaluate with model = 'TimesFM 3.0' for univariate time series evaluation in BigQuery.