Kairos-50M
Model directorymldi-lab/Kairos_50m50.1M params | Upstream checkpoint
Kairos-50M is the largest public checkpoint in the Kairos family in the public family, from the Ant Group and ShanghaiTech University Kairos release. It anchors the top of the Kairos size ladder and is the checkpoint to choose when you want the family's highest published capacity for zero-shot forecasting.
Architecturally it is the same adaptive encoder-decoder transformer as the smaller Kairos checkpoints: official sources describe an adaptive TSFM that varies tokenization granularity and instance-adaptive positional treatment based on the information density and temporal structure of each series. It is pretrained on the PreSTS corpus of 300B+ time points, and the larger parameter budget is meant to convert that broad pretraining into stronger coverage across heterogeneous domains.
Where to run this model
Compute availability, hardware requirements, and pricing depend on the execution provider.
Get model weights, deployment instructions, and current provider listings.
Hugging Face provider mappings checked 2026-10-06. No Inference Provider mapping was listed for this checkpoint.
Model Classification
Family
Kairos
Type
time series foundation model
Pretrained time-series model for zero-shot or few-shot forecasting workloads.
Resources
Training Data
PreSTS corpus with 300B+ time points, as documented by the official Kairos model cards and project page.
Recommended For
- • Adaptive zero-shot forecasting across heterogeneous series
- • Teams that want a mid-size open family with modern tokenization ideas
Strengths
- • Adaptive tokenization handles changing information density well
- • Clear parameter-size ladder from small to larger public checkpoints
Limitations
- • Newer family with a smaller production footprint than the most established lines
- • Focused on forecasting rather than general multi-task time-series tooling
Capabilities
Tags
Specifications
- Parameters
- 50.1M
- Architecture
- encoder-decoder transformer with adaptive patching and instance-adaptive rotary position encoding
- Context length
- n/a
- Max context
- n/a
- Minimum history
- n/a
- Recommended history
- n/a
- Input step
- n/a
- Required target series
- 1
- Temperature
- Ignored
- Top P
- Ignored
- Max output
- n/a
- License
- n/a