NeurIPS 2026 Competitions Announced

We are delighted to announce the competitions that have been accepted to the NeurIPS 2026 Competition Track. As in previous years, the selection process was highly competitive, with many outstanding submissions spanning a broad range of machine learning research areas and real-world applications. After a careful review process, we selected 16 exceptional competitions that reflect both emerging directions in AI research and long-standing challenges across science, engineering, and society, as listed below.

This year marks the tenth edition of the dedicated NeurIPS Competition Track. Whether you are participating in a NeurIPS competition for the first time or are a returning competitor, this year’s competition list offers exciting opportunities for researchers, practitioners, and students from a wide variety of backgrounds and expertise. We encourage everyone to explore the competitions and consider participating.

Competitions play a unique role in advancing machine learning research by bringing together communities around well-defined, impactful challenges for fair comparison of methods. They foster collaboration, encourage reproducible research, and provide shared benchmarks that accelerate progress across diverse domains.

As a participant, you will have the opportunity to engage with an active research community, develop new technical skills, and work on problems with meaningful real-world impact. Many past competitions have led to widely used benchmark datasets and open-source resources, scientific progress through the validation of impactful research directions and lasting collaborations. We hope this year’s competitions will continue that tradition.

Following the pattern from previous years, organizers of accepted competitions will have the opportunity to publish post-competition analyses either through the NeurIPS Datasets and Benchmarks Track or through a dedicated PMLR volume for NeurIPS competitions, providing a venue to document insights, benchmark results, and lessons learned from the challenges.

The NeurIPS 2026 program includes the following exciting competitions:

Scientific AI, Physics, and Engineering

• Learn2Design 2026: A Physics Experiment Design Competition for Gravitational Wave Detectors
• The Smart Buildings Challenge: Learning Foundation Models for Real-World Optimization at Scale
• Fusion Equilibrium Challenge: Inferring & Transferring Magnetic Geometry without Magnetic Diagnostics
• NeurIPS 2026 RealPDE Competition: Scientific ML for Real-World Physical Systems
• FAIR Universe Weak Lensing ML Uncertainty Challenge on Distribution Shifts in Cosmology

Healthcare, Biology, and Humanitarian Applications

• Synthetic Humanitarian Surveys at Scale (SimulacraBench): A Closed Evaluation via Unreleased UN Microdata
• AMP Challenge: A Benchmark for Generative AI in Antimicrobial Peptide Design
• Speech Accessibility Project Challenge 2
• The Virtual Embryo Challenge: Generative Modeling of Embryogenesis Across Space, Scale and Time

Foundation Models, Reasoning, and AI Evaluation

• The Steerability Challenge
• Predictive AI Evaluation Competition
• AIMO Interpretability Challenge
• QuantiPhy Challenge: Quantitative Physical Reasoning for VLMs

Robotics, Agents, and Embodied AI

• RoboSynChallenge: Mastering Real-World Dexterity via Generalizing Synthesized Manipulation Skills
• RoCo-Spring: The Robust Correspondence Challenge
• Agenthon 2026: Verifiable AI for Quantitative Finance

We are deeply grateful to all of the reviewers and members of the Competition Track Program Committee who dedicated their time and expertise to evaluating this year’s proposals. We congratulate the organizers of the accepted competitions and sincerely thank everyone who submitted a proposal for contributing such high-quality ideas.

If you have any questions or suggestions regarding the Competition Track, please feel free to contact competition-chairs@neurips.cc. 

We look forward to the outcomes of this year’s competitions and to welcoming organizers and participants at NeurIPS 2026. 

NeurIPS 2026 Competition Track Chairs
Jes Frellsen, Shou-de Lin, Annika Reinke