A researcher claims to have developed Jev, a new frontier AI model trained using RLCD methodology that is significantly faster and cheaper than existing models, with free output tokens and optimized for decision-making tasks.
Diogo Almeida and team released Jev, a new frontier AI model trained using RLCD methodology, which outperforms Astra, Fable, and Opus on radiofrequency engineering tasks with claimed 20-200x faster performance.
TypeSafe released Jev, an early-access model that answers structured questions with calibrated probabilities rather than generating text. The model uses RLCD training instead of RLHF to optimize for reliable predictions, costs $0.042 per million input tokens, and was tested on 24 Norwegian documents about salmon farming and tax policy.
TypeSafe AI develops Machine Native Intelligence, a production-focused AI system optimized for narrow, inspectable decisions in software rather than general-purpose responses. It introduces RLCD (Reinforcement Learning from Calibrated Decisions), a post-training approach that returns probabilistic decisions with reliable confidence scores instead of generated text, addressing limitations of RLHF like hallucinations and mode collapse.