On June 3, main blockchain and AI organizations, EigenLayer, Cyber, Sentient amongst them, introduced the institution of the CAIBA – Crypto AI Benchmark Alliance. The aim of this community-led initiative is to set clear requirements for AI fashions and crypto brokers’ analysis.
The members of the Alliance are 14 organizations: Alchemy, Cyber, EigenLayer, Goldsky, IOSG, LazAI, Magic Newton, Metis, MyShell, OpenGradient, RootData, Sentient, Surf, Thirdweb.
These corporations are offering datasets, instruments, and specialised data to develop a benchmarking framework. Each benchmark, printed brazenly on developer platforms, will include outlined duties, reference options, and grading scripts.
A benchmark for CAIA – Crypto AI Brokers – is the primary launch of the Alliance and is presently out there. The benchmark supplies correct answering protocol and token-related questions, mapping out multi-step duties and instruments like block explorers and APIs to finish duties. There are duties included in CAIA, associated to tokenomics, on-chain evaluation, challenge analysis, and transaction workflows.
New benchmarks are actually underway, and the Alliance welcomes new contributors. Builders, researchers, and protocols are inspired to take part by submitting fashions for analysis or proposing new benchmark duties.
Picture: CAIBA
