Artificial intelligence is radically transforming the healthcare sector, and Google Research has just launched a major initiative to accelerate this revolution. The MedGemma Impact Challenge, hosted on Kaggle since January 13, 2026, invites developers, researchers, and teams from around the world to create human-centered AI applications in the medical field. With a prize pool of $100,000 and access to cutting-edge open-source models.
The Challenge: Building Tomorrow’s Medical AI
The MedGemma Impact Challenge is based on a simple but crucial observation. Many clinical environments require AI systems capable of operating locally, without permanent Internet connectivity, while respecting strict data privacy requirements. Participants must develop concrete applications using MedGemma 1.5 and other models from Google’s Health AI Developer Foundations (HAI-DEF) program.
Application areas are vast: high-dimensional medical imaging (CT, MRI, histopathology), extracting structured data from medical reports, anatomical localization, and multimodal workflows combining text, images, and speech.
Each submission must include three elements:
- a demonstration video of maximum 3 minutes,
- a technical document of maximum 3 pages,
- and reproducible source code published via Kaggle Writeups.
Participants can submit and revise their project multiple times before the deadline, thus encouraging an iterative approach rather than a single delivery.
Practical Info & MedGemma Hackathon Calendar
The challenge takes place entirely online on the Kaggle platform, allowing worldwide participation without geographical constraints.
Key Dates
- Launch: January 13, 2026
- Final submission deadline: February 24, 2026
- Evaluation period: March 2026
Prizes and Rewards
- Main prizes: $75,000 distributed across 4 positions
- Additional special prizes: $25,000 for thematic categories (AI agent-based workflows, innovative fine-tuning adaptations, edge AI deployment on local devices)
- Total prize pool: $100,000
Who Can Participate in the MedGemma Impact Challenge?
The challenge is open to everyone: developers, researchers, students, healthcare professionals with technical skills. Teams can participate, but the maximum size and exact team composition modalities are not explicitly detailed.
No geographical restrictions are mentioned, making the competition truly global. Submissions must be made via Kaggle Writeups, which requires a Kaggle account (free). MedGemma 1.5 models and other HAI-DEF resources are available for free via Hugging Face and Google Cloud Vertex AI.
Evaluation Criteria & MedGemma Hackathon Rules
Submissions will be evaluated by a jury composed of members from Google Research, Google DeepMind, and Google’s health AI teams according to five major criteria:
- Effective use of HAI-DEF models: relevant exploitation of MedGemma, MedSigLIP, MedASR or other models from the program
- Importance of the problem addressed: clinical relevance and potential impact of the use case
- Possible real impact: deployment feasibility in real healthcare environments
- Technical feasibility: soundness of approach, code quality, reproducibility
- Execution and communication quality: demonstration clarity, technical documentation
Resources & Official Links
Why Participate in This MedGemma Hackathon?
Beyond the $100,000 in rewards, the MedGemma Impact Challenge offers a unique opportunity to learn and innovate at the frontier of medical AI. Participants get free access to cutting-edge models developed by Google Research, benefit from Kaggle’s collaborative environment with shared notebooks and approaches, and can iterate publicly on their solutions.
It’s also a chance to get noticed by Google’s AI teams and join a global network of healthcare AI developers. The most promising projects could attract the attention of healthcare organizations, medtech startups, or lead to academic collaborations. For students and professionals in career transition, it’s a high-level portfolio project demonstrating multimodal AI skills applied to a critical domain.
