Granite Timeseries Ensemble

IBM · released Oct 22, 2026

Open-weight (permissive)time series

Openness taxonomy

access, licence, and disclosure — never a single open/closed binary

Weights availability: Ensemble configuration is downloadable; four member checkpoints are downloaded at runtime.
Licence: unknown
Licence category: Permissive licence stated for the ensemble recipe and default Granite members; specific licence not named.
Commercial use allowed: unknown
Derivatives allowed: unknown
Redistribution allowed: unknown
Field-of-use restrictions: unknown
Training code availability: unknown
Inference code availability: unknown
Architecture disclosure: The ensemble recipes and four member architectures are described; configuration stores ensemble settings rather than merged weights.
Training data disclosure: unknown

openness confidence

Source states the ensemble recipe and default Granite members are permissively licensed.
Source states the recipe loads from Hugging Face and downloads member checkpoints at runtime.

Deployability

raw values as extracted — free text, never inferred from branding

Self-hosting: unknown
On-premise: unknown
Edge deployability: unknown
API availability: unknown
Regions available: unknown

Benchmarks

as reported — vendor-supplied unless otherwise noted

No benchmark data on this release yet.

National-security pathway & evidence

from the assessed source item, when one is linked

Impact: lowConfidence: mediumNat-sec: noneprimary

Evidence limitations: Official Hugging Face Blog material is a single primary source; the supplied excerpt is truncated and provides no evaluation results or named licence.

Ensemble forecasting with IBM Granite time-series models

Linked research papers

from paper_model_links

No linked research papers yet.