SAVRN
Search Contact SAVRN

Open-weight model · Time series forecasting

Toto-2.0-22m

by Datadog Datadog/Toto-2.0-22m

Toto (Time Series Optimized Transformer for Observability) is a family of time series foundation models for multivariate forecasting developed by Datadog.

Parameters22M
Context
Weights87.7 MB
Licenseapache-2.0
AccessOpen weights
Monthly Downloads125k

Runs On

What it takes to serve Toto-2.0-22m (22M parameters): the memory its weights need at each precision, and the cheapest way to rent enough data-center GPUs to hold them.

PrecisionWeightsMemory neededCheapest setupPer hourAlso fits
16-bit 0.0 GB 0.1 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
8-bit 0.0 GB 0.0 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00
4-bit 0.0 GB 0.0 GB 1x MI300X (192 GB)
Vultr
$1.85 1x H100 $1.99 · 1x MI325X $2.00

Memory is the weights at that precision plus 20% for the runtime and a short context; a long context needs more. Prices are the lowest on-demand hourly rates in the SAVRN Index, read Sep 18, 2026.

Model Card

By Datadog, published under apache-2.0, revision 685e4ae3e2be.

Toto (Time Series Optimized Transformer for Observability) is a family of time series foundation models for multivariate forecasting developed by Datadog. Toto 2.0 is the current generation, featuring u-μP-scaled transformers ranging from 4m to 2.5B parameters, all trained from a single recipe. Forecast quality improves reliably with parameter count across the family.

The family sets a new state of the art on three forecasting benchmarks: BOOM, our observability benchmark; GIFT-Eval, the standard general-purpose benchmark; and the recent contamination-resistant TIME benchmark.

Performance

Quick Start

Inference code is available on GitHub.

Installation

pip install toto-models

Inference Example

Read the full model card (457 words)

Identity and Version

Repository
Datadog/Toto-2.0-22m
Publisher
Datadog
Task
Time series forecasting
Modality
Time series
Library
pytorch
Parameters
22M parameters
Languages
Not stated by the source
Revision
685e4ae3e2be8d8998025e53dd98e7fdcb296a89
First published
2026-04-14
Last updated
2026-06-04

Files and Weights

7 files, 88.8 MB in total. The weights are 1 file totalling 87.7 MB in safetensors.

Weights1 file · 87.7 MB
Configuration1 file · 593 B
Documentation1 file · 7.2 KB
Other3 files · 1.1 MB
Repository1 file · 1.7 KB
Every file
FileTypeSizeSHA-256
model.safetensorsWeights87.7 MB 9cd503d82df3
config.jsonConfiguration593 B
README.mdDocumentation7.2 KB
assets/architecture.pngOther436.7 KB 973196289f60
assets/pareto.pngOther302.2 KB 756a05902735
figures/architecture.pngOther396.1 KB d17a304f234c
.gitattributesRepository1.7 KB

License and Download

License
apache-2.0
Access
Open weights, no gate
Download size
87.7 MB
Download from Datadog

Released by Datadog through its official repository on Hugging Face. Read the license.

Built From

Evaluations

Each result is shown as reported, with the conditions its reporter stated. None is a SAVRN measurement. A comparison lines two results up only when their configuration, unit and setup are all stated and identical.

BenchmarkConditionsResultReported byRevisionDate
BOOM Task time-series-forecastingMetric CRPSComparison conditions not established 0.363 Datadog
Publisher reported
Evaluated revision not stated
BOOM Task time-series-forecastingMetric MASEComparison conditions not established 0.601 Datadog
Publisher reported
Evaluated revision not stated
GIFT-Eval Task time-series-forecastingMetric CRPSComparison conditions not established 0.496 Datadog
Publisher reported
Evaluated revision not stated
GIFT-Eval Task time-series-forecastingMetric MASEComparison conditions not established 0.719 Datadog
Publisher reported
Evaluated revision not stated
TIME Task time-series-forecastingMetric CRPSComparison conditions not established 0.556 Datadog
Publisher reported
Evaluated revision not stated
TIME Task time-series-forecastingMetric MASEComparison conditions not established 0.668 Datadog
Publisher reported
Evaluated revision not stated

Memory Requirements

PrecisionWeights in memory
As published87.7 MB
16-bit0.0 GB
8-bit0.0 GB
4-bit0.0 GB

Weights only, from the published parameter count; the key-value cache and runtime add to this.

Compare Toto-2.0-22m

Questions About Toto-2.0-22m

How much GPU memory does Toto-2.0-22m need?

About 0.1 GB at 16-bit and 0 GB at 4-bit: the weights (22M parameters) plus a working margin. A long context needs more.

What is the cheapest GPU to run Toto-2.0-22m on?

At 16-bit, 1x MI300X from $1.85 an hour; at 4-bit, 1x MI300X from $1.85 an hour, at the lowest on-demand prices the SAVRN Index lists.

Can I use Toto-2.0-22m commercially?

Yes. Toto-2.0-22m is released under Apache License 2.0. The Apache License 2.0 is a permissive open-source license. It permits commercial use, modification and redistribution. It requires keeping the license and copyright notices and any NOTICE file, stating significant changes, and it includes an express patent grant from contributors.

Similar Models

Model · Time series forecasting

chronos-bolt-mini

Autogluon

Update Feb 14, 2025: Chronos-Bolt models are now available on Amazon SageMaker JumpStart! Check out the tutorial notebook to learn how to deploy Chronos endpoints for production use in a few lines of code. Chronos-Bolt is a family of pretrained time series forecasting models which can be used for zero-shot forecasting. It is based on the T5 encoder-decoder architecture and has been trained on nearly 100 billion time series observations. It chunks the historical time series context into patches of multiple observations, which are then input into the encoder. The decoder then uses these representations to directly generate quantile forecasts across multiple future steps—a method known as…

Open weights apache-2.0 21M parameters

Model · Time series forecasting

chronos-bolt-mini

Amazon

Update Feb 14, 2025: Chronos-Bolt models are now available on Amazon SageMaker JumpStart! Check out the tutorial notebook to learn how to deploy Chronos endpoints for production use in a few lines of code. Chronos-Bolt is a family of pretrained time series forecasting models which can be used for zero-shot forecasting. It is based on the T5 encoder-decoder architecture and has been trained on nearly 100 billion time series observations. It chunks the historical time series context into patches of multiple observations, which are then input into the encoder. The decoder then uses these representations to directly generate quantile forecasts across multiple future steps—a method known as…

Open weights apache-2.0 21M parameters chronos-forecasting

Model · Time series forecasting

chronos-t5-mini

Autogluon

Update Feb 14, 2025: Chronos-Bolt & original Chronos models are now available on Amazon SageMaker JumpStart! Check out the tutorial notebook to learn how to deploy Chronos endpoints for production use in a few lines of code. Update Nov 27, 2024: We have released Chronos-Bolt models that are more accurate (5% lower error), up to 250 times faster and 20 times more memory-efficient than the original Chronos models of the same size. Check out the new models here. Chronos is a family of pretrained time series forecasting models based on language model architectures. A time series is transformed into a sequence of tokens via scaling and quantization, and a language model is trained on these…

Open weights apache-2.0 20M parameters transformers

Model · Time series forecasting

chronos-t5-mini

Amazon

Update Feb 14, 2025: Chronos-Bolt & original Chronos models are now available on Amazon SageMaker JumpStart! Check out the tutorial notebook to learn how to deploy Chronos endpoints for production use in a few lines of code. Update Nov 27, 2024: We have released Chronos-Bolt models that are more accurate (5% lower error), up to 250 times faster and 20 times more memory-efficient than the original Chronos models of the same size. Check out the new models here. Chronos is a family of pretrained time series forecasting models based on language model architectures. A time series is transformed into a sequence of tokens via scaling and quantization, and a language model is trained on these…

Open weights apache-2.0 20M parameters chronos-forecasting

Model · Time series forecasting

Kronos-small

ShiYu

Kronos is the first open-source foundation model for financial candlesticks (K-lines), trained on data from over 45 global exchanges. It is designed to handle the unique, high-noise characteristics of financial data. Kronos is a family of decoder-only foundation models, pre-trained specifically for the "language" of financial markets—K-line sequences. It leverages a novel two-stage framework: 1. A specialized tokenizer first quantizes continuous, multi-dimensional K-line data (OHLCV) into hierarchical discrete tokens. 2. A large, autoregressive Transformer is then pre-trained on these tokens, enabling it to serve as a unified model for diverse quantitative tasks. The success of large-scale…

Open weights mit 25M parameters

Model · Time series forecasting

chronos-2-small

Autogluon

This is the small variant of the Chronos-2 model with 28M parameters. For usage and details on the Chronos-2 model, please refer to https://huggingface.co/autogluon/chronos-2. If you find Chronos-2 useful for your research, please consider citing the associated paper

Open weights apache-2.0 28M parameters chronos-forecasting