Project AURIONAI Use-Cases Research & Innovation
time series

TimesFM 2.5

TimesFM version behind BigQuery AI.FORECAST that 3.0 is planned to replace (implied).

GoogleGoogleUSATimesFMFamily
Google
Levelmodel
Classtime series
Parameters200M
Contextnot in the notes
Released2 Sep 2025
Opennessopen weights permissive
Runs onsingle datacentre GPU, CPU server
Statusactive

Benchmark Results

vendor-reportedindependent
GIFT-Evaltime series
zero-shot foundation-model CRPS rank, lower is bettervendor-reported
25.8
02040

Score in mean CRPS rank. Lower is better.

Toto 2.0 report (Datadog)Jun 2026 arxiv.org
zero-shotindependentno test-data leakage
0.705
00.71.4

Score in MASE (geo-mean, normalised to Seasonal Naive). Lower is better.

Leaderboard rebuilt from the published raw results, 143 entries covering all 97 configurations9 Oct 2026 huggingface.co
zero-shot; rank is average CRPS rankindependentno test-data leakage
0.49
00.51

Score in CRPS (geo-mean, normalised to Seasonal Naive). Lower is better.

rank 66 of 143, Leaderboard rebuilt from the published raw results, 143 entries covering all 97 configurations9 Oct 2026 huggingface.co
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

Lineage

successor
TimesFM 1.0 to TimesFM 2.0: successor, medium confidenceTimesFM 2.0 to TimesFM 2.5: successor, medium confidenceTimesFM 2.5 to TimesFM 3.0: successor, high confidenceTimesFM 1.0TimesFM 1.0GoogleTimesFM 2.0TimesFM 2.0GoogleTimesFM 2.5TimesFM 2.5Google, Sep 2025TimesFM 3.0TimesFM 3.0Google, Aug 2026
GoogleTimesFM 2.0 is the parent (successor)medium confidencenotes
GoogleTimesFM 3.0 is a child (successor)high confidencenotes
Where this comes from
Parent and child relations from model cards, Hugging Face base_model fields and the genealogy tables of the research notes. The tree shows two generations each way, 4 models. Confidence and evidence are the notes' own.

Industry Use Cases

Industrial
Industrial forecasting (load, throughput, demand)
Retail and Logistics
Demand forecasting

R&D Topics

Time-series foundation modelsproduced by

Identity

Developer
Google
Family
TimesFM
Released
2 Sep 2025 sourcesregistry date
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
dense transformer Decoder-only
Total parameters
200M sources
Active parameters
–
Checkpoint parameters
–
Note
timesfm-2.5-200m-pytorch
Context length
not in the notes
Input
time series
Output
time series
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open weights permissive
Licence
Apache License 2.0 sources
Access routes
open weights download, cloud marketplace
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
single datacentre GPU, CPU server
Footprint
Cloud (BigQuery univariate forecasting path, unverified) or local. Hugging Face update 2 Oct 2025.
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Sources

  1. arxiv.org/pdf/2605.20119 https://arxiv.org/pdf/2605.20119
  2. thenewstack.io/google-timesfm-3-multivariate-forecasting https://thenewstack.io/google-timesfm-3-multivariate-forecasting/
  3. TimesFM vs Chronos vs Moirai: The Leader's Weights Ban Production beri.net https://www.beri.net/article/timesfm-vs-chronos-vs-moirai-time-series-foundation-model-licence-bigquery-sagemaker-cost-2026
  4. huggingface.co/api/models/google/timesfm-2.5-200m-pytorch https://huggingface.co/api/models/google/timesfm-2.5-200m-pytorch
  5. google/timesfm-2.5-200m-pytorch , Hugging Face huggingface.co https://huggingface.co/google/timesfm-2.5-200m-pytorch
  6. github.com/google-research/timesfm https://github.com/google-research/timesfm

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "TimesFM 2.5", https://aurion.semai.id/models/google-timesfm-2-5/, records of October 2026.