Discover our 2026 challenges

The SoccerNet Challenges return for their 6th edition, with brand-new tasks and upgraded classics that push the frontier of sports AI. The competition runs until April 24, 2026, giving you time to craft breakthrough solutions and climb the leaderboard.

Stay engaged with updates, standings, and discussions on Discord, and dive into our YouTube tutorials and highlights for inspiration.

Whether you’re tackling a new task or refining last year’s work, join us in shaping the future of sports AI. SoccerNet Challenges 2026 await—are you in?

CHAT WITH US ON DISCORD

Spiideo SoccerNet SynLoc 2026

📢 1️⃣ First Edition of the Spiideo SoccerNet SynLoc Challenge! 🚀
Single-Frame World-Coordinate Athlete Detection & Localization (with Synthetic Data)

In collaboration with Spiideo, this challenge tests single-frame world-coordinate athlete localization: detect all players and predict their pitch positions in meters. Using a mix of synthetic + real broadcast data with precise calibration, models must bridge domain gaps across stadiums and cameras. Scored with mAP-LocSim for both detection and localization accuracy.

🗂️Data / 💻Development Kit / 📚Rules / 📊Eval: [test] [challenge]

🤝Sponsor: Spiideo / 💰Prize: USD 1000

Ball Action Anticipation

🔉2️⃣ Ball Action Anticipation Challenge! 🚀

Ball action anticipation involves anticipating both the timing and type of ball-related actions that will occur within the next 5 seconds. 30-second gameplay videos are provided before each anticipation window, and the actions span across 10 classes in our dataset. Each action is marked by a single timestamp.

🗂️ Data & Development Kit / 📊Eval: [test] [challenge]

🤝Sponsor: interested in sponsoring? If yes, please reach out! / 💰Prize: TBA

VQA Visual Question Answering

📢 3️⃣ VQA Visual Question Answering Challenge!

We are thrilled to launch the SoccerNet Challenge 2026 - VQA, inspired by our work "Multi-Agent System for Comprehensive Soccer Understanding" (MM'25). This edition introduces Foundation Models and Agentic AI to tackle General Knowledge reasoning in football. Participants will answer close-ended questions across 14 diverse soccer understanding tasks spanning text, image, and video modalities, marking a significant step toward Generative AI in Football.

🗂️Data / 💻Development Kit / 📚Rules / 📊Eval: [test] [challenge]

🤝Sponsor: KNQ Technology / 💰Prize: USD 1,000

Player-Centric Ball Action Spotting

📢 3️⃣ Player-Centric Ball Action Spotting Challenge!

A new high-level computer vision challenge that extends the SoccerNet Action Spotting series by focusing not only on what and when an event happens, but also on who performs it. The Player-Centric Ball Action Spotting task aims to identify both the temporal occurrence and the player responsible for each ball-related action across full soccer matches.

🗂️Data / 💻Development Kit / 📊Eval: [test] [challenge]

🤝Sponsor: Footovision / 💰Prize: USD 1,000

More challenges will be announced soon...