# THOR: A Versatile Foundation Model for Earth Observation Climate and Society Applications

arXiv: https://arxiv.org/abs/2601.16011 · HTML: https://mlsanity.com/paper/2601.16011/

## Key facts

- Cited by 7 papers in the map (data through 14 Oct 2026).
- Builds on 21 of its 37 references that are in the map.
- First posted on arXiv 22 Jan 2026.

## Builds on (most cited first, 21 of 21)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Masked Autoencoders Are Scalable Vision Learners](https://mlsanity.com/paper/2111.06377/) (2021)
- [Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture](https://mlsanity.com/paper/2301.08243/) (2023)
- [SatMAE: Pre-training Transformers for Temporal and Multi-Spectral Satellite Imagery](https://mlsanity.com/paper/2207.08051/) (2022)
- [Scale-MAE: A Scale-Aware Masked Autoencoder for Multiscale Geospatial Representation Learning](https://mlsanity.com/paper/2212.14532/) (2022)
- [CROMA: Remote Sensing Representations with Contrastive Radar-Optical Masked Autoencoders](https://mlsanity.com/paper/2311.00566/) (2023)
- [FlexiViT: One Model for All Patch Sizes](https://mlsanity.com/paper/2212.08013/) (2022)
- [Prithvi-EO-2.0: A Versatile Multi-Temporal Foundation Model for Earth Observation Applications](https://mlsanity.com/paper/2412.02732/) (2024)
- [Self-supervised Learning in Remote Sensing: A Review](https://mlsanity.com/paper/2206.13188/) (2022)
- [TerraMind: Large-Scale Generative Multimodality for Earth Observation](https://mlsanity.com/paper/2504.11171/) (2025)
- [Mission Critical -- Satellite Data is a Distinct Modality in Machine Learning](https://mlsanity.com/paper/2402.01444/) (2024)
- [MMEarth: Exploring Multi-Modal Pretext Tasks For Geospatial Representation Learning](https://mlsanity.com/paper/2405.02771/) (2024)
- [AnySat: One Earth Observation Model for Many Resolutions, Scales, and Modalities](https://mlsanity.com/paper/2412.14123/) (2024)
- [PANGAEA: A Global and Inclusive Benchmark for Geospatial Foundation Models](https://mlsanity.com/paper/2412.04204/) (2024)
- [Towards a Unified Copernicus Foundation Model for Earth Vision](https://mlsanity.com/paper/2503.11849/) (2025)
- [Masked Autoencoders for Microscopy are Scalable Learners of Cellular Biology](https://mlsanity.com/paper/2404.10242/) (2024)
- [USat: A Unified Self-Supervised Encoder for Multi-Sensor Satellite Imagery](https://mlsanity.com/paper/2312.02199/) (2023)
- [Multi-Label Guided Soft Contrastive Learning for Efficient Earth Observation Pretraining](https://mlsanity.com/paper/2405.20462/) (2024)
- [FlexiMo: A Flexible Remote Sensing Foundation Model](https://mlsanity.com/paper/2503.23844/) (2025)
- [Domain-specific optimization and diverse evaluation of self-supervised models for histopathology](https://mlsanity.com/paper/2310.13259/) (2023)
- [On Pretraining Data Diversity for Self-Supervised Learning](https://mlsanity.com/paper/2403.13808/) (2024)

## Built on by (most cited first, 7 of 7)

- [TESSERA v2: Scaling Pixel-wise Earth Foundation Models](https://mlsanity.com/paper/2607.03949/) (2026)
- [Any Model, Any Place, Any Time: Get Remote Sensing Foundation Model Embeddings On Demand](https://mlsanity.com/paper/2602.23678/) (2026)
- [SpectralEarth-FM: Bringing Hyperspectral Imagery into Multimodal Earth Observation Pretraining](https://mlsanity.com/paper/2605.21075/) (2026)
- [Spatial Representation Learning Beyond Pixels: Unifying Raster Data and Vector Semantics for Human-Centric Geospatial Foundation Models](https://mlsanity.com/paper/2606.02374/) (2026)
- [Multi-Modal, Multi-Temporal Geospatial Foundation Models Improve Global Above-Ground Biomass Estimation](https://mlsanity.com/paper/2608.04792/) (2026)
- [Now We Know? A Systematic Comparison of TerraMind and THOR](https://mlsanity.com/paper/2607.18504/) (2026)
- [Earth Observation Foundation Models for Terrestrial Ecohydrology: From Representation Learning to Process Inference](https://mlsanity.com/paper/2608.15282/) (2026)

## Abstract (arXiv)

Current Earth observation foundation models are architecturally rigid, struggle with heterogeneous sensors and are constrained to fixed patch sizes. This limits their deployment in real-world scenarios requiring flexible computeaccuracy trade-offs. We propose THOR, a "computeadaptive" foundation model that solves both input heterogeneity and deployment rigidity. THOR is the first architecture to unify data from Copernicus Sentinel-1, -2, and -3 (OLCI & SLSTR) satellites, processing their native 10 m to 1000 m resolutions in a single model. We pre-train THOR with a novel randomized patch and input image size strategy. This allows a single set of pre-trained weights to be deployed at inference with any patch size, enabling a dynamic trade-off between computational cost and feature resolution without retraining. We pre-train THOR on THOR Pretrain, a new, large-scale multi-sensor dataset and demonstrate state-of-the-art performance on downstream benchmarks, particularly in data-limited regimes like the PANGAEA 10% split, validating that THOR's flexible feature generation excels for diverse climate and society applications.