Blockchain-based smart contracts offer a promising approach to securely manage and monetize machine learning (ML) models within the banking and finance sector. This paper introduces a protocol leveraging Ethereum blockchain technology to create trustless contracts for the evaluation and exchange of ML models, enabling financial institutions to access high-quality predictive models without direct reliance on third-party trust. Financial firms can submit datasets and evaluation functions to the blockchain, establishing automated contracts that incentivize data scientists to submit optimized ML models for specific financial applications like credit scoring, fraud detection, or risk assessment. Our approach leverages cryptographic validation to ensure model performance transparency and accuracy, eliminating counterparty risk and establishing a decentralized marketplace for AI solutions in finance. This blockchain-driven protocol not only facilitates secure model exchange but also promotes competitive pricing and the efficient use of computational resources, thus enhancing AI accessibility in finance while mitigating operational and cybersecurity risks.