A
Meta has just released Tribe v2, an open-source version of its groundbreaking foundation model designed to forecast how the human brain responds to virtually any image or sound.
Built on Meta’s Algonauts 2025 award-winning architecture, Tribe v2 draws from over 500 hours of functional MRI (fMRI) recordings collected from more than 700 participants. The model effectively creates a “digital twin” of neural activity — allowing researchers to predict brain responses for new subjects, languages, and tasks without any prior analysis.
This release marks a major step forward in computational neuroscience and AI. By mapping sensory input directly to neural responses, Tribe v2 could unlock new possibilities in brain–computer interfaces, cognitive research, and personalized AI models that adapt to how we think and perceive the world.
Meta’s decision to open-source the system underscores growing momentum toward transparent, human-centered AI research — positioning Tribe v2 as a pivotal tool for scientists aiming to bridge the gap between artificial intelligence and the human mind.
