Snorkel AI Open Benchmarks Grants
Snorkel AI Open Benchmarks Grants give in kind data services, research time and credits to teams building open AI benchmarks. Rolling applications.
The Snorkel AI Open Benchmarks Grants program funds open source datasets, benchmarks and evaluation artifacts for frontier AI systems. Snorkel backs the program with a $3M commitment. Applications are reviewed on a rolling basis and there is no fixed deadline.
The support is in kind. No cash is paid. Grantees can get access to Snorkel's Expert Data as a Service offering, dedicated time from its research and engineering teams, and compute and platform credits from selected partner platforms. Any stated grant amount is the estimated value of these services.
The call is for researchers, labs and engineers who build benchmarks for the next wave of AI capabilities. Snorkel names three areas it cares about: environment complexity, autonomy horizon and output complexity. It also welcomes independent directions from the research community.
To apply, fill in the short form on the apply page and add supporting material if you have it. A steering committee of academic and industry researchers reviews the proposals. Selected teams then work with Snorkel and partner research teams.
Grantees keep the IP in what they build, but must publish it under a permissive open license, such as MIT or Apache 2.0 for code and CC BY 4.0 or CC0 for data. Papers should acknowledge Snorkel's support. You also need approval from your institution where its policies require it.