TiRex
Model directoryNX-AI/TiRex35M params | NX-AI Community License | Upstream checkpoint
TiRex is NX-AI's xLSTM-based zero-shot forecasting model, a compact foundation forecaster that deliberately steps away from the Transformer backbone shared by most models in this catalog. The official card highlights strong short- and long-horizon performance together with support for both point and quantile forecasts, making it a capable general-purpose zero-shot option rather than a narrow specialist.
Architecturally it is a 35M-parameter xLSTM forecasting model, which gives it the recurrent-style in-context generalization behaviour that distinguishes it from attention-only designs. Its pretraining provenance is documented at a high level: the official card lists autogluon/chronos_datasets and Salesforce/GiftEvalPretrain as the main sources, without publishing a more exact mixture breakdown.
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.
- Retrocast · The Forecasting Company
managed service · Source checked 2026-10-06
Forecast through the managed API using tirex. The provider controls the deployed variant and request limits.
- Author-hosted demo · Hugging Face Spaces
hosted demo · Source checked 2026-10-06
Try the author’s interactive forecasting demo. Computation runs on the Space, not on your device.
- Community forecasting playground
hosted demo · Source checked 2026-10-06
The published loader names this checkpoint. Space was sleeping when checked; it may require a wake-up. Hosted computation, not browser inference.
- sktime TServe
self-hosted runtime · Family / provider variant · Source checked 2026-10-06
Run a model-serving container on your own hardware or compute account. Select the catalog checkpoint and matching runtime image.
Hugging Face provider mappings checked 2026-10-06. No Inference Provider mapping was listed for this checkpoint.
Model Classification
Family
TiRex
Type
time series foundation model
Pretrained time-series model for zero-shot or few-shot forecasting workloads.
Resources
Training Data
Official card lists autogluon/chronos_datasets and Salesforce/GiftEvalPretrain as the main sources, without publishing a more exact mixture breakdown.
Recommended For
- • Compact zero-shot forecasting when you want an alternative to Transformers
- • Short- and long-horizon forecasting with probabilistic outputs
Strengths
- • xLSTM backbone differentiates it from the dominant Transformer families
- • Competitive capacity-to-quality ratio
Limitations
- • Smaller ecosystem than the biggest public TSFM families
- • Less familiar operationally if your team only benchmarks Transformer-based models
Capabilities
Tags
Specifications
- Parameters
- 35M
- Architecture
- xLSTM-based forecasting model
- 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
- NX-AI Community License