Built on Ai2’s planet-scale Earth AI models, OlmoEarth Platform unlocks access to real-time earth data and insights to enable organizations to build solutions in service of people and the planet 

Ai2 (The Allen Institute for AI) today announced the OlmoEarth Platform, the first open, end-to-end solution that transforms global satellite and sensor data into real-time earth insight. At its core is OlmoEarth, Ai2’s state-of-the-art multimodal foundation model family, trained on millions of Earth observations spanning roughly ten terabytes of data.

Built on these advanced models, OlmoEarth Platform brings cutting-edge AI to governments, NGOs, and local communities—now making it possible to monitor deforestation, assess crop health, and predict wildfire risk without deep AI expertise or specialized infrastructure.

Every second, sensors and satellites capture enormous amounts of data about our changing planet, revealing where forests are disappearing, which crops are under stress, and where climate risks are growing. Until now, much of that insight has been siloed and locked behind proprietary systems.

The OlmoEarth Platform changes that by providing an open, scalable workflow that spans data collection, labeling, model training, inference, and deployment. Powered by the OlmoEarth models, which learn from diverse multimodal, multi-temporal data – combining radar, optical, and environmental signals to achieve best-in-class performance across environmental domains – OlmoEarth Platform empowers organizations everywhere to generate and act on advanced Earth intelligence.

“The challenges facing our planet today demand solutions that are not only powerful but are accessible to all,” said Ali Farhadi, CEO of Ai2. “With OlmoEarth, we’re making Earth AI accessible to those working on the front lines. It’s built to help governments, NGOs, and communities that wouldn’t otherwise have access to AI. It enables them to use it through an adaptable, and open platform—allowing them to see change as it happens and empowering them to respond quickly in a time of need.”

Powerful, Fine-Tunable Foundation Models

The OlmoEarth model family includes a range of sizes—from lightweight models designed for fast, low-cost inference to larger, high-capacity models for precision mapping and analysis. All share a unified, easily fine-tunable architecture that enables temporal and spatial understanding of immense amounts of Earth data.

Across benchmarks, OlmoEarth outperforms larger commercial and academic models, achieving state-of-the-art results on key tasks like crop mapping, mangrove extent estimation, and live fuel moisture prediction. These models can be customized for new local conditions with relatively modest datasets, putting world-class environmental AI within reach of any organization.

Unlike conventional systems that require separate models for each task, OlmoEarth’s generalizable design enables a single adaptable model to be fine-tuned across multiple use cases, reducing costs, time, and barriers to action.

Global Deployments Show Early Impact

Across the globe, early adopters are already demonstrating OlmoEarth Platform’s real-world potential:

• In Kenya, the International Food Policy Research Institute (IFPRI) is using OlmoEarth to create countywide crop maps informed by local field data, helping officials to anticipate challenges, target seed and fertilizer distribution, and strengthen food security strategies.
• In the Amazon, the Amazon Conservation Association is detecting drivers of deforestation in near real-time across Peru, Bolivia, Colombia, and Brazil, supporting targeted enforcement and recovery efforts of local partners, including government agencies and indigenous communities.
• In early testing with Global Mangrove Watch, OlmoEarth shows accuracy of about 97%, which could cut data processing time in half. Faster updates enable conservationists and governments to detect loss events sooner, plan restoration work, and target interventions where they’re needed most.
• In wildfire readiness, partners are using the OlmoEarth Platform to map live fuel moisture at scale, helping planners focus prevention and response where risk is rising. Ai2 is also working with NASA’s Jet Propulsion Laboratory (JPL) as part of the AI Collaborative on Wildfire to inform next-generation fire risk modeling.

“An annual map that arrives with a few years’ delay doesn’t really help you as a protected area manager,” said Lammert Hilarides from Wetlands International. “For deforestation alerts, timing is even more critical. You need to know as early as possible if mangroves are being cleared or degraded, so you can act quickly.”

Tools That Bring AI to Everyone

OlmoEarth Platform includes a suite of integrated tools for building, deploying, and sharing geospatial AI development:

• OlmoEarth multimodal foundation models power OlmoEarth—trained on massive datasets of Earth observations, available in a range of sizes and customizable to local needs and requirements.
• OlmoEarth Studio provides an intuitive workspace for creating datasets, fine-tuning models, and collaborating across teams and organizations.
• OlmoEarth Viewer enables anyone to explore model-generated maps through an interactive web app, turning complex data into visual, actionable insights.
• OlmoEarth Run manages large-scale compute jobs for fine-tuning, inference, and dataset generation, ensuring results are reproducible and scalable.
• OlmoEarth APIs enable developers to connect the platform directly to existing tools and workflows, bringing planetary intelligence to any application.

“The greatest threats facing our planet cannot be solved through isolated efforts. This platform's holistic approach to Earth AI will enable deeper collaboration and drive scientific innovation,” said Patrick Beukema, lead researcher on the OlmoEarth team. “By working together, across disciplines and organizations, on a shared, open foundation, we can accelerate progress and deliver solutions to the world's most urgent challenges.”

The OlmoEarth Platform is available now, and Ai2 is currently accepting requests for additional partnerships and collaborations. As with all Ai2’s fully open AI models, Ai2 is releasing code, documentation, and example models through OlmoEarth Projects, a public GitHub repository showcasing how anyone can train and deploy fine-tuned models using the OlmoEarth foundation models. The commitment is simple: insights that stay fresh, models that adapt, and applications that anyone can use, customize, and change—all in service of people and the planet.

About Ai2

Ai2 is a Seattle-based non-profit AI research institute with the mission of building breakthrough AI to solve the world’s biggest problems. Founded in 2014 by the late Paul G. Allen, Ai2 develops foundational AI research and innovative new applications that deliver real-world impact through large-scale open models, open data, robotics, conservation platforms, and more. Ai2 champions true openness through initiatives like OLMo, the world’s first truly open language model framework, Molmo, a family of open state-of-the-art multimodal AI models, and Tulu, the first application of fully open post-training recipes to the largest open-weight models. These solutions empower researchers, engineers, and tech leaders to participate in the creation of state-of-the-art AI and to directly benefit from the many ways it can advance critical fields like medicine, scientific research, climate science, and conservation efforts. For more information, visit allenai.org.

View source version on businesswire.com: https://www.businesswire.com/news/home/20251104384267/en/

“The challenges facing our planet today demand solutions that are not only powerful but are accessible to all. With OlmoEarth, we’re making Earth AI accessible to those working on the front lines." - Ali Farhadi, CEO of Ai2