Eligibility and key facts
Eligibility to be confirmed The summary below does not confirm your eligibility. Check eligible regions, age, language, and other requirements.
- Who may be interested
- Researchers, Developers, University students
- Team size
- Individuals or teams of up to 5 may participate. Requires a Kaggle account and the older of 18 or the local age of majority, unless the sponsor agrees with appropriate guardian consent. Residents of Crimea, the so-called DNR or LNR, Cuba, Iran or North Korea, and persons or entity representatives under U.S. export controls or sanctions, cannot enter. Local laws also apply. Competition-entity personnel may participate subject to internal policy but cannot win prizes.
- Location
- Online on Kaggle
- Eligible regions
- To be confirmed; online does not imply worldwide eligibility
- Entry fee
- Free
- Prizes and support
- US$77,000 total: US$59,000 for the main leaderboard and US$18,000 for the efficiency track. Eligible submissions can qualify for both.
All deadlines
Original date: Entry deadline | October 15, 2026 - Entry Deadline. You must accept the competition rules before this date in order to compete. All deadlines are at 11:59 PM UTC. | Time zone: UTC
This deadline has passed and cannot be added.
Original date: Team join / merger deadline | October 15, 2026 - Team Merger Deadline. This is the last day participants may join or merge teams. All deadlines are at 11:59 PM UTC. | Time zone: UTC
This deadline has passed and cannot be added.
Original date: Final submission deadline | October 22, 2026 - Final Submission Deadline. All deadlines are at 11:59 PM UTC. | Time zone: UTC
Original date: Winners' requirements deadline | November 5, 2026 - Winners' Requirement Deadline. This is the deadline for winners to submit to the host/Kaggle their training code, video and method description. All deadlines are at 11:59 PM UTC. | Time zone: UTC | Only applies to winners; not a new-entry or general submission deadline.
Already registered? Check your submission deadline. Saving an event does not register you.
About this opportunity
Train multimodal models using knee MRI studies and paired radiology reports to detect twelve important abnormalities. The main leaderboard uses macro-averaged ROC AUC, with a separate track rewarding prediction quality and runtime efficiency. Individuals or teams of up to 5 may participate. Requires a Kaggle account and the older of 18 or the local age of majority, unless the sponsor agrees with appropriate guardian consent. Residents of Crimea, the so-called DNR or LNR, Cuba, Iran or North Korea, and persons or entity representatives under U.S. export controls or sanctions, cannot enter. Local laws also apply. Competition-entity personnel may participate subject to internal policy but cannot win prizes. Rules explicitly require no purchase to enter or win; participants' own equipment and compute costs are separate.
Rules and deliverables
Participation rules
- Individuals or teams of up to 5 may participate. Requires a Kaggle account and the older of 18 or the local age of majority, unless the sponsor agrees with appropriate guardian consent. Residents of Crimea, the so-called DNR or LNR, Cuba, Iran or North Korea, and persons or entity representatives under U.S. export controls or sanctions, cannot enter. Local laws also apply. Competition-entity personnel may participate subject to internal policy but cannot win prizes.
- Up to five submissions per day, two final submissions and five members per team.
- Submit via Notebooks with CPU or GPU runtime at most nine hours and internet disabled; freely and publicly available external data and pretrained models are allowed.
- Data access uses the MIRA license. The winner license is labelled CC-BY-NC 4.0, but the rules contain conflicting commercial-use wording; clarify it with the organizer before entering.
- Efficiency awards require a final-evaluation submission above the sample benchmark on the private leaderboard and rank submissions using runtime and predictive performance.
- Worldwide, subject to official eligibility, residency, sanctions and local-law restrictions
What to prepare
- A Notebook producing submission.csv with confidence scores for twelve abnormalities for each StudyInstanceUID.
- Winners provide training and inference code, weights, environment documentation, a method description and short video, and publicly release code, weights and a model for distribution and validation.
Sources
- Source platform
- Kaggle
- Source domain
- www.kaggle.com
The organizer’s published schedule and rules take precedence.