What is ClimateGPT? 

ClimateGPT is a powerful family of open–source AI models, developed by Erasmus.AI. It’s designed to support smarter, better climate-related decision-making. The model enables researchers, policymakers, and business leaders to generate trustworthy insights faster. Since the release of 2.0 version, ClimateGPT has already been downloaded more than 10,000 times on Hugging Face. 

The models are trained on a vast (100B+) dataset of carefully selected, reliable climate-related texts. These include scientific papers, news articles, and reports from international development organisations, treaty organisations, and the broad NGO community (e.g., the World Bank, United Nations, NASA, etc.). Erasmus.AI added structured data from more than 50 external databases to the 2.0 version. This allows ClimateGPT to perform complex analyses, such as tracking carbon emissions, identifying renewable energy trends, and forecasting weather patterns. 

With this rich dataset, ClimateGPT generates clear, context-aware answers to a wide range of climate-related questions, from scientific fundamentals to advanced modelling and mitigation strategies. 

How open-source AI improves climate decision-making

Open-source AI provides (much needed) transparency and trust. It also enables researchers, governments, and organisations worldwide to share data, models, and innovations. By opening up the code, developers and climate scientists can iterate quickly, spot flaws, and build on each other’s work. This obviously has a very positive effect on the speed with which innovations take place. Another way in which open-source AI improves decision-making is that it allows smaller, less wealthy countries to also take part in the conversation, which – to stay in line with our own lingo – vastly expands the “data set”. In short: open-source AI brings speed, scale, and inclusiveness to the fight against climate change, all while making sure the decision-making tools themselves are trustworthy and fair.  

Our contribution to the project

INFO works closely with Erasmus.AI to develop new ClimateGPT applications. Currently, we’re building: 

• Climate Policy Agent: the user interface of ClimateGPT, which started as a basic prompt-engineering tool, is evolving into a smart digital agent. Users can already explore climate topics from economic, political, and social perspectives. Soon, ClimateGPT will support their entire workflow, from generating reports to running simulations automatically. 
• ClimateMCP: we’re implementing the Model Context Protocol (MCP), an emerging, open standard that connects AI to external tools and data sources. This creates a more relevant, seamless experience when using ClimateGPT. It gives LLMs the right context at the right time, making conversations more relevant, more insightful, and – in the end – more useful. 

For INFO, this partnership goes beyond technology. It’s an opportunity to design and build digital products that embed Responsible AI into everyday business processes, from scenario planning and risk analysis to sustainability reporting. 

Why open–source AI matters

By making both source code and training data publicly available, researchers, developers, and organisations worldwide can contribute to its quality, transparency, and usability. 

INFO firmly believes in this mission and in open-source product development. Open access to science, data and digital solutions is paramount for a more sustainable future.  

Get involved 

Would you like to know more about how AI can create value and support your organisations goals? Get in touch with our team and explore the (many!) options.