$HYPERAGENTS
Supporting Meta AI Research on pump.fun
100% of the fees are redirected to Meta AI Research GitHub. We believe HyperAgents will change AI forever.
Contract Address
EtdfHNFGGhCAW6Met9EBAhBr7jQMKtASaYC6DoHwpumpWhat is HyperAgents?
HyperAgents is a groundbreaking research from Meta AI that introduces self-referential agents— AI systems that can improve themselves by modifying their own code and behavior.
Unlike traditional self-improving AI that relies on fixed meta-level mechanisms, HyperAgents integrates a task agent (which solves problems) and a meta agent (which modifies itself) into a single editable program. This enables metacognitive self-modification.
Self-Accelerating Progress
Can potentially support self-accelerating progress on any computable task, improving not just solutions but the improvement process itself.
Cross-Domain Transfer
Meta-level improvements transfer and accumulate across domains including coding, robotics, paper review, and Olympiad math.
Darwin Gödel Machine
Built on DGM-H (Darwin Gödel Machine - Hyperagents), which generates and evaluates self-modified variants through a growing archive.
Open-Ended AI
Offers a glimpse of AI systems that don't just search for better solutions, but continually improve how they search for improvement.
The Research
Meta AI Research
March 23, 2026
Authors
Jenny Zhang, Bingchen Zhao, Wannan Yang, Jakob Foerster, Jeff Clune, Minqi Jiang, Sam Devlin, Tatiana Shavrina
University of British Columbia, Vector Institute, University of Edinburgh, New York University, Canada CIFAR AI Chair, FAIR at Meta, Meta Superintelligence Labs
View on GitHubWhy $HYPERAGENTS?
This is more than a memecoin. It's a community-driven initiative to support the future of AI research.
100%
of fees redirected to Meta AI Research GitHub
Community
Join believers in self-improving AI on X
Future
Supporting research that could change AI forever