Project AURIONAI Use-Cases Research & Innovation
Microsoft AI
weather

Aurora

Microsoft Research Earth-system foundation model trained on more than one million hours of geophysical data, with fine-tunes for air quality, waves and cyclones. Code MIT; weight licence not stated on Azure page.

Map of tropical cyclone track predictions by Aurora compared with official forecasts.
Levelmodel
Classweather
Parameters1.3B
Contextnot in the notes
Released21 May 2025
Opennessopen source
Runs onmulti-GPU node, single datacentre GPU
Statusactive

Benchmark Results

maintainer-run
Z500 day 10, deterministic, Aurora (oper.) vs Analysismaintainer-run
668.4
0450900

Score in RMSE m2/s2. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
Z500 day 5, deterministic, Aurora (oper.) vs Analysismaintainer-run
262.8
0450900

Score in RMSE m2/s2. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
Z500 day 3, deterministic, Aurora (oper.) vs Analysismaintainer-run
119.8
0450900

Score in RMSE m2/s2. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
Z500 day 1, deterministic, Aurora (oper.) vs Analysismaintainer-run
38.1
0450900

Score in RMSE m2/s2. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
T850 day 10, deterministic, Aurora (oper.) vs Analysismaintainer-run
2.979
024

Score in RMSE K. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
T850 day 5, deterministic, Aurora (oper.) vs Analysismaintainer-run
1.501
024

Score in RMSE K. Lower is better.

year 2022, 240x121 grid10 Oct 2026 researchsites.withgoogle.com
Where this comes from
Every score names who ran it: vendor-reported, maintainer-run or independent. A rank is shown with the population it was ranked in (gap rule 3).

R&D Topics

Identity

Developer
Microsoft AI
Family
not in the notes
Released
21 May 2025 sourcesannouncement page
Weights published
–
Record checked
10 Oct 2026
Where this comes from
Developer and dates from the research notes. Release date is the vendor's public announcement; weights date is when open weights were published.

Architecture and Size

Architecture
dense transformer 1.3B atmosphere foundation model; about 11 km
Total parameters
1.3B sources
Active parameters
–
Checkpoint parameters
–
Context length
not in the notes
Input
sensor
Output
sensor
Where this comes from
Parameter counts as the developer states them. Where the checkpoint holds a different total, both are shown (register A16).

Openness and Access

Openness
open source
Licence
MIT License sources
Access routes
open weights download
Where this comes from
The licence file governs over registry tags and aggregator labels.

Deployment

Runs on
multi-GPU node, single datacentre GPU
Footprint
PyPI microsoft-aurora 2.0.1, 4 Aug 2026; tagged Experimental on Azure
Runtimes
not in the notes
Where this comes from
Deployment tiers and runtimes as named in the notes. Footprint is the notes' figure, not a measurement by AURION.

Indonesia

Indonesian support
not stated
Regional languages
not in the notes
Jakarta hosting
unknown
Languages
not in the notes
Where this comes from
Indonesia fields are carried on every model; unknown is an allowed value.

Lineage

The notes record no lineage for this model.

Industry Use Cases

No industry use case links this model.

Sources

  1. Aurora - Microsoft Foundry Labs labs.ai.azure.com https://labs.ai.azure.com/innovations/aurora/
  2. A foundation model for the Earth system - Nature nature.com https://www.nature.com/articles/s41586-025-09005-y
  3. GitHub - microsoft/aurora: Implementation of the Aurora model for Earth system forecasting github.com https://github.com/microsoft/aurora
  4. pypi.org/pypi/microsoft-aurora/json https://pypi.org/pypi/microsoft-aurora/json
  5. huggingface.co/api/models/microsoft/aurora https://huggingface.co/api/models/microsoft/aurora

To cite this page: Project AURION, AI Use-Cases Research & Innovation, "Aurora", https://aurion.semai.id/models/microsoft-aurora/, records of October 2026.