Kronos Small
Model directoryNeoQuasar/Kronos-small24.7M params | MIT | Upstream checkpoint
Kronos Small is the mid-size checkpoint in the NeoQuasar Kronos family, sitting between the lightweight Mini variant and the higher-capacity Base model. It is the natural step up when Kronos Mini leaves accuracy on the table but the full Base checkpoint is more than a workload needs.
It uses a larger tokenizer vocabulary and more transformer capacity than the Mini variant, which is what improves accuracy on complex multi-asset financial forecasting tasks while keeping the same overall design — a GPT-style decoder over tokens produced by the family's financial candlestick tokenizer. Like the rest of the family it is designed for OHLCV candlestick data; the official examples require at least OHLC columns, with volume and amount remaining optional. The official Kronos model cards describe pretraining on over 12B K-line records from 45 global exchanges.
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.
- 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
Kronos
Type
time series foundation model
Pretrained time-series model for zero-shot or few-shot forecasting workloads.
Resources
Training Data
Official Kronos model cards describe pretraining on over 12B K-line records from 45 global exchanges.
Recommended For
- • Financial time-series forecasting with OHLCV candlestick data
- • Multi-asset price prediction across equities, forex, and crypto
Strengths
- • Purpose-built tokenizer for financial price movement patterns
- • Native OHLCV support with synthetic fallback for univariate series
Limitations
- • Domain-specific to financial data — not a general-purpose TSFM
- • Best fit when open, high, low, and close covariates are provided explicitly; the univariate fallback is a compatibility path, not the ideal contract
- • Limited public documentation and benchmark coverage compared to major TSFM families
Capabilities
Tags
Specifications
- Parameters
- 24.7M
- Architecture
- GPT-style decoder with financial candlestick tokenizer
- 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
- MIT