IBM releases SOTA Granite Time Series PatchTST-FM-r2 model with commercial-friendly license

Summary

IBM has released Granite Time Series PatchTST-FM-r2, a roughly 385-million-parameter foundation model for zero-shot forecasting across demand, prices, energy loads, traffic, telemetry, and other time series. The model uses conformer blocks combining multi-head self-attention with temporal convolution, 50% overlapping patches with Hamming-window weighting, overlap-and-add forecasting, contexts up to 8,192 steps, and a 99-quantile prediction head for probabilistic outputs and imputation support. On GIFT-Eval, it ranks second for both CRPS and MASE among replicable zero-shot models as of September 8, 2026, and first in that category among permissively licensed models; including pretrained models, it ranks third for CRPS and fourth for MASE. IBM provides open weights, architecture, inference code, and reproducibility code under Apache 2.0 and OpenMDW 1.0. The release also describes Python use and streaming integration through Apache Flink on Confluent Cloud.

Takeaways

1. 1

   The model’s documented pretraining corpus includes selected GiftEvalPretrain datasets, modified-periodic-kernel KernelSynth data, TSMixup data restricted to datasets outside GIFT-Eval, and about 500,000 synthetic CauKer sequences of length 4,096.

2. 2

   Conformer blocks use alternating temporal-convolution kernel sizes of 5 and 3 in the repeating pattern {5, 5, 3, 3}, allowing convolution to capture shorter-term interactions while attention focuses on longer-range relationships.

3. 3

   PatchTST-FM-r2 is dual licensed under Apache 2.0 and OpenMDW 1.0, and its architecture implementation is backward-compatible with PatchTST-FM-r1 checkpoints.