Tsinghua University logo

Sundial (Base 128M)

online
thuml/sundial-base-128m

128M params | 3K context | $0.00025 per forecast

Sundial is THUML's native generative TSFM for continuous-valued forecasting. The official card presents it as a 128M-parameter decoder-only model trained with TimeFlow Loss, built to generate multiple plausible futures without relying on discrete tokenization. It is one of the most directly probabilistic models in the hosted catalog and a strong option when distribution quality matters as much as point accuracy.

Model Classification

Family

Sundial

Type

time series foundation model

Pretrained time-series model exposed on TSFM.ai for zero-shot or few-shot forecasting workloads.

Training Data

TimeBench / 1032B time points, with the model card listing THUML UTSD, LOTSA, and Chronos datasets as core upstream sources.

Recommended For

  • Probabilistic forecasting where distribution quality matters
  • Teams that want native continuous generative forecasting rather than tokenization

Strengths

  • Designed around probabilistic future generation
  • Native continuous-valued modeling rather than discrete token bins

Limitations

  • Less suited to broad non-forecasting tasks
  • Generative probabilistic focus can be overkill for simple point-forecast-only use cases

Capabilities

forecastingprobabilistic-forecastingquantile-forecastingzero-shot

Tags

sundialgenerativeprobabilisticthuml

Specifications

Parameters
128M
Architecture
causal decoder-only transformer with TimeFlow loss / flow matching
Context length
2,880
Max context
2,880
Minimum history
n/a
Recommended history
n/a
Input step
n/a
Required target series
1
Temperature
Ignored
Top P
Ignored
Max output
1,024
Avg latency
n/a
Uptime
n/a
Plan limits
1,000 rpm free · 1,000,000 rpm with billing
Accelerator
T4
Regions
Virginia, US
License
n/a

Pricing

Per forecast
$0.00025

Performance

Average latency
n/a
Availability
n/a
Plan limits
1,000 rpm free · 1,000,000 rpm with billing