---
id: engines/terramind
about:
  label: TerraMind
  description: >-
    Open generative foundation model for Earth observation from IBM and the
    European Space Agency: it reads radar, optical, elevation and land-cover
    layers and can generate any one of them from the others.
  tags:
    - satellite
    - foundation model
    - embeddings
    - sentinel
    - generative
tool:
  form: open model
  group: Earth observation and AI
  homepage: 'https://ibm.github.io/terramind/'
  licence: Apache-2.0 (code and weights)
  cost: free
domains:
  - places
  - ecology
open:
  source: open
  data: unknown
  api: unknown
  participation: open
  licence: Apache-2.0
  note: >-
    Weights for tiny, small, base and large models are on Hugging Face;
    fine-tuning runs through IBM's TerraTorch toolkit.
relations:
  - predicate: alternative_to
    target: engines/tessera
  - predicate: alternative_to
    target: engines/alphaearth-prithvi
provenance:
  origin: agent
  sources:
    - title: TerraMind (project page)
      url: 'https://ibm.github.io/terramind/'
    - title: IBM/terramind (README)
      url: 'https://github.com/IBM/terramind'
    - title: ibm-esa-geospatial/TerraMind-1.0-base (Hugging Face model card)
      url: 'https://huggingface.co/ibm-esa-geospatial/TerraMind-1.0-base'
    - title: >-
        IBM and ESA open-source TerraMind, the best performing generative AI
        model for Earth observation
      url: 'https://research.ibm.com/blog/terramind-esa-earth-observation-model'
  created: '2026-10-21'
  reviewed: '2026-10-21'
  reviewed_by: 'Claude, from the sources listed'
  status: draft
---
TerraMind summarises satellite imagery as numbers, like other Earth observation foundation models, and it can also generate one kind of layer from another: a radar image or a land-use map from an ordinary Sentinel-2 picture, for instance.[^1,2] A generated layer is the model's guess at what that layer would show, and the model card warns against reading it as a reconstruction.[^3] On the project page, a Sentinel-2 scene of Singapore becomes a radar image in which ships and runways show and the clouds are gone.[^1]

It was built by IBM, the European Space Agency's Φ-lab and the FAST-EO project, and released on Hugging Face on 22 April 2025.[^1,4] KP Labs, the Jülich Supercomputing Centre, where it was trained, and the German Aerospace Center (DLR) also took part.[^1,4] The paper appeared at ICCV 2025, a computer vision conference.[^1,2] It is a separate model from the Prithvi models IBM makes with NASA ([AlphaEarth / Prithvi](/engines/alphaearth-prithvi/)); IBM's announcement names both as part of its work applying AI to the planet, and says NASA experts helped validate TerraMind.[^4]

It was trained on TerraMesh, nine million matched samples from around the world in nine kinds of data.[^4] These include Sentinel-1 radar, Sentinel-2 optical imagery, land use and land cover, a vegetation index, elevation, coordinates and short text descriptions.[^1,4] As an input it accepts Sentinel-2 at two processing levels, Sentinel-1 in two radar formats, an elevation model or a plain colour image, and it copes when some of these are missing.[^3]

Run as an encoder on your own imagery, the base model gives 768 numbers for each 16 by 16 pixel patch of a 224 by 224 pixel input, averaged across the input layers by default.[^3] Fine-tuning goes through IBM's TerraTorch toolkit, with example settings for flood mapping, burn scars and crop types, and multi-date data goes through a time-series wrapper in that toolkit.[^2] An option called Thinking-in-Modalities has the model first imagine a missing layer, land cover by default, and then use it to improve its answer.[^1,2]

Code and weights are under Apache 2.0 and come in tiny, small, base and large sizes.[^2,3]